Master’s Thesis

(1)

Master’s Thesis

The Effect of the Social Dimension in Quantified Self Applications on App Usage Continuation Intention and the Moderating Role of User Personality Type

Author Friederike Behrends

Student number 11754036

Date of submission June 24, 2022 Version being submitted Final

Study program MSc. Business Administration – Digital Marketing track Name of institution University of Amsterdam, Amsterdam Business School EBEC approval number EC 20220506070532

(2)

Statement of originality

This document is written by student Friederike Behrends who declares to take full responsibility for the contents of this document. I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it. The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.

(3)

This Master’s thesis is written as part of the Business Administration Master’s at the University of Amsterdam. The main purpose of this research is to understand the effect of the social dimension within Quantified Self applications on users’ app usage continuation intention, as well as to examine whether individual differences in personality traits affect this relationship.

I would like to thank a number of people, who helped me during the thesis process:

First, I would like to thank my supervisor Dr Michael Etter; not only for his valuable feedback throughout the last few months, but also for the interesting discussions during meetings on the topic of QS applications and for providing a clear sense of direction. Thank you very much.

I would also like to thank the second reader of this thesis in advance for your time and effort reading this thesis. Besides, my gratitude goes to everyone who participated in this study and who therefore contributed to new knowledge in the field of Quantified Self applications and social psychology. Lastly, I am extremely grateful for the constant support of my family and friends during my Master’s. I could not have taken this journey without you.

(8)

Abstract

This study focuses on the relationship between three elements of the social dimension in Quantified Self applications, namely recognition, reciprocal benefits, and network exposure, and users’ app usage continuation intention. Furthermore, it tests whether the Big Five personality traits extraversion, agreeableness, conscientiousness, neuroticism, and openness to experience have an impact on these relationships. It is hypothesized that extraversion and neuroticism have a significant positive effect on the relationship between network exposure and QS app usage continuation intention, while agreeableness and conscientiousness are expected to positively moderate the relationship between reciprocal benefits and the outcome variable, whereas openness to experience is expected to positively moderate the relationship between recognition and the outcome variable. To test these hypotheses, a survey is conducted among current, past, and nonusers of QS applications (N=182). In the first section of the survey, a personality assessment is conducted, followed by a visual and text-based scenario displaying elements of the social dimension in QS applications. Furthermore, it is assessed how respondents perceive recognition, reciprocal benefits, and network exposure, followed by statements pertaining to their app usage continuation intention based on the scenario provided.

The results indicate that while all three elements of the social dimension have a significant positive effect on app usage continuation intention, none of the personality traits have a moderating effect; except for conscientiousness, which has a significant negative effect on the relationship between reciprocal benefits and app usage continuation intention. Furthermore, results do not differ significantly by gender.

(9)

1. Introduction

In modern society, humans are facing an increase in work-related stress and a growing pressure of work intensification and competition caused by globalization, the growth of innovation technologies, and continuously increasing workloads (Foy et al., 2019). In conjunction with the rising demands of modern society comes a lack of leisure time, impeding people to balance out their demanding day-to-day lives, e.g., by engaging in physical activity (Chen & Pu, 2014).

Yet, scarce leisure time is not the only factor to blame for a lack of participation in physical exercise in today’s world. Humans have a high tendency to favor immediate benefits over long- term rewards, a bias called hyperbolic discounting, or present bias, causing people to frequently neglect actions which could be beneficial in the long-term (Hamari & Koivisto, 2015).

New technologies have emerged in recent years attempting to ‘motivate people by restructuring relatively long-term goals by providing the users with short-term goals, activities, rewards, and social support’ (Hamari & Koivisto, 2015). An example of these technologies are health and fitness applications, which have gained popularity in the last few years. According to industry statistics, sports and fitness applications have increased at an annual growth rate of 150% in recent years (Lin et al., 2020). As one of the leading applications in the industry, Nike Run Club registered a 45% increase in downloads from 2019 to 2020 (Shaban, 2021). Besides a surge in total downloads, these applications registered a significant growth in daily active users, increasing by 24% from Q1 to Q2 of 2020 alone (Ang, 2020).

Combining this growing demand with new technological advances, health and fitness applications have been introducing a variety of features attempting to provide short-term rewards, which motivate people to engage in physical activity. In June 2021, GPS running and cycling app Strava introduced its newest feature, a competition-based in-app challenge, allowing users to create virtual events in which they are competing against up to 24 friends

(10)

breakthrough feature is one of the consequences of a growing demand for mobile health and fitness applications, fueled by the COVID-19 pandemic and an increased need to not only stay active, but concurrently connected with each other.

A vast majority of health and fitness applications utilize the Quantified Self (QS), which Lupton (2016) defines as ‘practices in which people knowingly and purposively collect information about themselves, which they then review and consider applying in their lives’

(Lupton, 2016). The QS movement, which allows humans to self-track their behavior, actions, and other quantifiable data, including their diets, step count, heart rate, and distance travelled (Brinson & Rutherford, 2020), has been of interest to researchers for many years. The main technologies facilitating user self-quantification are 1) wearables and 2) self-tracking applications, of which the latter are the focus of this thesis.

While the statistics in the introductory section demonstrate a positive trend of a continuously increasing application user base, app usage discontinuation is a rising concern.

Despite high download and usage rates, a growing number of application users either decides to stop actively using an app, or to uninstall the application after a short period of time (Cho, 2016). With 74% of QS app users abandoning the app within two weeks of usage (Stragier et al., 2016), the churn rate of self-tracking apps is considerably high. As health-tracking technologies are facing a growing challenge of usage discontinuance (Huang et al., 2019), skepticism concerning the effectiveness of QS applications in positively contributing to users’

lives is rising (Clawson et al., 2015). Understanding the motives for usage continuation intention is crucial not only from a managerial perspective, but also from a behavioral perspective: If the factors leading to app usage continuation and abandonment are known, not only can marketing professionals implement the necessary adaptive steps by particularly incorporating in-app elements which users value; the perceived usefulness created for the end user increases simultaneously, potentially decreasing the churn rate of QS applications as a

(11)

result (Bhattacherjee, 2001). It has been found that significant user benefits resulting from health and fitness application usage start appearing after sufficient time, i.e., 4 to 6 months of continued usage (Marcolino et al., 2018); hence, prolonging active app usage past this point is in the interest of businesses to attain a high level of perceived user benefits.

Numerous factors have been found to impact app usage continuation of QS applications, namely ease of use (Zhou et al., 2017), design features (Vaghefi & Tulu, 2019), and motivational features including social features and gamification elements (Hamari et al., 2018). As one of the fundamental components within QS applications, the social dimension encompasses interactions with an established network of friends, family members, and other athletes. The social dimension has the potential of fostering user engagement, while being a central source of usage continuation due to the inherent psychological need of human beings for experiencing relatedness and connection (Hamari & Koivisto, 2015; Zhang, 2008). Based on a preliminary literature review within the domain of QS applications, it could be determined that although the social dimension in QS applications has been extensively explored in previous studies, limited research has been conducted on app usage continuation intention as an outcome variable, with the study by Hamari & Koivisto (2015) being the exception. Scholars have called for additional research on the topic, as previous studies have primarily addressed alternative outcomes of the social dimension in self-tracking applications, e.g., the consequences for well- being (James et al., 2019), changes in user motivation and behavior (Barratt, 2017), user attitudes and exercise continuance (Hamari & Koivisto, 2015), and communication within the social circle (Lomborg & Frandsen, 2015). Besides, several researchers investigating the social dimension in QS applications conducted their study within the boundary of a single QS app, such as Strava. While this is not a drawback per se, a study providing insights with a higher applicability and generalizability is required.

(12)

To measure the effect of the social dimension on users’ behavior, Hamari & Koivisto (2015) introduced recognition, reciprocal benefits, and network exposure as main elements of the social dimension. Hamari & Koivisto (2015) studied the effect of these variables on people’s willingness to exercise and long-term usage intention of exercise gamification technologies and found significant positive effects; yet, they based their research on current users of exercise gamification services, as opposed to past users or nonusers. As valuable information can be gained from all user groups, including current users, past users, and nonusers, the aim of the current study is to confirm the findings of Hamari & Koivisto (2015) in a different setting.

Yet, as acknowledged by Alshawmar et al. (2022), people’s individual characteristics may influence how the social dimension in QS applications is perceived. The authors studied the effect of personality traits on fitness app usage continuation intention and concluded that conscientiousness and neuroticism personality traits positively influence perceived usefulness of the fitness application, which in turn increases usage continuation intention. Moreover, earlier studies suggest that personality traits have the potential of affecting people’s technology use and social media use (Amichai-Hamburger & Ben-Artzi, 2003; Guadagno et al., 2008; Unal et al., 2017). Thus, the current research aims to further extend previous studies by analyzing whether differences in personality traits influence people’s perception of the social dimension in QS applications, and specifically of the variables recognition, reciprocal benefits, and network exposure introduced by Hamari & Koivisto (2015).

Based on the present knowledge gap, this thesis addresses the following research question: What is the effect of the social dimension in QS applications on app usage continuation intention? To answer this research question, quantitative research is conducted among current, past, and nonusers of QS applications. The aim is hereby to confirm that recognition, reciprocal benefits, and network exposure have a significant positive effect on QS

(13)

app usage continuation intention, and to examine whether differences in personality traits moderate these relationships.

This thesis contributes to theory and practice in the following ways: It extends the existing knowledge base by analyzing the impact of the social dimension in QS applications on app usage continuation intention and validates whether the findings of Hamari & Koivisto (2015) regarding the positive effects of recognition, reciprocal benefits, and network exposure on app usage continuation intention influence all user groups. The study also examines the moderating effects of personality traits on these relationships, hereby adding on to the existing body of research. Furthermore, marketeers, consumers, and corporations can benefit from these insights, as the study analyzes whether elements of the social dimension have the ability to influence consumer behavior, and whether users’ personality traits have an effect on this relationship. This newly generated knowledge will aid app developers, corporations, and marketeers in adapting application design and marketing strategies according to the profile of the end user.

This research is structured as follows: In the subsequent chapter, literature pertaining to the social dimension in QS applications is reviewed, encompassing literature of the three dimensions recognition, reciprocal benefits, and network exposure. Furthermore, a review of relevant literature in the field of personality psychology within the context of digital social applications and general technology usage is conducted. Chapter 3 comprises of the conceptual framework, displaying the independent variables, the dependent variable, and the moderating variables in a visual manner. Thereafter, the research methods are discussed in chapter 4, followed by the results section in chapter 5, presenting the main findings of this study. The results of the analysis are further discussed in chapter 6, and include theoretical implications, managerial implications, limitations, and directions for future research, followed by the

(14)

2. Literature review

This chapter provides an overview of existing literature within the field of social technologies, QS applications, and personality psychology. First, the social dimension in QS applications is discussed, further comprising of recognition, reciprocal benefits, and network exposure. Next, literature on the topic of the personality traits extraversion, agreeableness, conscientiousness, neuroticism, and openness to experience and their effect on technology and social media usage is discussed, followed by an overview of literature on the role of gender.

2.1 Social dimension in QS applications

The central features of QS applications are inherently user-centric, as personal data is being collected on individual user statistics, health data, and activity records (Brinson & Rutherford, 2020). Yet, in recent years, health and fitness applications have taken on an additional role by increasingly incorporating social features, allowing users to interact with an established network of friends, family members, and other athletes. In line with Whelan & Clohessy’s (2021) work, the terminology social dimension is used to describe in-app features designed to foster social relationships between users, incorporating elements of social influence within the boundaries of QS applications. The social dimension in information technologies has the potential of fostering user engagement, while being a central source of usage continuation due to the inherent psychological need of human beings for experiencing relatedness and connection (Hamari & Koivisto, 2015; Zhang, 2008).

As recognized by Lomborg & Frandsen (2015) and congruent with self-determination theory, humans strive to meet three fundamental psychological needs when using media;

namely 1) autonomy, 2) competence, and 3) relatedness to others (Deci & Ryan, 1985;

Lomborg & Frandsen, 2015), indicating that social interaction with community members is a primary human need. Furthermore, Vallerand & Losier (1999) found that ‘individuals are

(15)

intrinsically motivated to move toward situations and experiences that will satisfy these basic needs’ (Vallerand & Losier, 1999), presenting a major opportunity for businesses intending to unlock the potential of social interactions. These theories on the nature of human psychology provide reasoning for the modern tendency of humans to engage in a continuous search for meaningful social interactions in online media, including social media (Lomborg, 2014;

Lomborg & Frandsen, 2015). Furthermore, the importance of the social dimension within QS applications shows parallels to dynamics within sports communities, in which social support and peer recognition are central constituents, particularly visible in competitions and winner celebrations (Lomborg & Frandsen, 2015).

It has been recognized that the social dimension has a significant effect on technology use (Beldad & Hegner, 2017; Vannoy & Palvia, 2010). Previous research acknowledged the importance of the social dimension within self-tracking applications, as multiple studies confirm that social influence is one of the main drivers impacting QS app usage, linked to increased usage satisfaction and higher engagement (Hamari & Koivisto, 2015; James et al., 2019; Whelan & Clohessy, 2021). Furthermore, studies confirm that individuals who view themselves as being an integral part of a health and fitness community show higher levels of participation within this group, e.g., by engaging with in-group members to a greater extent, while also fostering an individual’s sensitivity and interest for their own health (Dessart &

Duclou, 2019). Social motives have been found to be particularily important for experienced users of online fitness communities as opposed to novice users (Stragier et al., 2016), emphasizing that the social dimension is a central factor for customer retention in highly loyal customer groups. Moreover, in a broader context, users’ self-promotion on social networking sites may contribute to positive consequences in a non-digital context, including positive changes in well-being, self-esteem, and life satisfaction (Islam et al., 2019).

(16)

As one of the few studies investigating app usage continuation as a dependent variable, Beldad & Hegner (2017) investigated the social dimension in the context of health and fitness applications, however, their focus lies entirely on the social dimension in the non-digital environment as opposed to within applications, which is the focus of this thesis. To extensively analyze the social dimension in the digital context, the following variables will be studied: 1) Recognition, 2) reciprocal benefits, and 3) network exposure. These variables were incorporated in previous research as elements of the social dimension of health and fitness applications, however, relating them to different outcome variables, namely well-being (Whelan & Clohessy, 2021), willingness to exercise, and gamification usage intention (Hamari

& Koivisto, 2015).

2.1.1 Recognition

In the context of QS applications, recognition pertains to social responses including feedback and acknowledgement, which users receive based on their actions and behavior (Whelan &

Clohessy, 2021), and which typically results from actions of the user’s self-promotion. Within the context of Internet-based technologies, recognition is a commonly used interaction feature allowing individuals to obtain social feedback in response to their undertaken activies (Cheung et al., 2011; Cheung & Lee, 2010; Hamari & Koivisto, 2015; Lin, 2008). Peer recognition generally involves a positive response directed at the action of another user; in the context of social networks, this oftentimes includes feedback provided in the form of likes, thumbs up, and other symbols related to generating a positive reaction (Whelan & Clohessy, 2021). QS applications have incorporated a similar strategy with the aim of providing users with positive recognition for their physical activities, which are generated by the user’s social network;

examples of which are game design elements, including ‘badges, medals, and trophies for personal bests and noteworthy performances’ (Whelan & Clohessy, 2021), along with liking

(17)

and commenting features. Running app Strava has adopted the ‘kudos’ feature, which is similar to liking a post on social networks, and which provides athletes with a quick in-app recognition on their physical achievement, e.g., when reaching a milestone (Whelan & Clohessy, 2021).

Recognition of individual performance is a basic requirement for fulfilling the basic human need of experiencing relatedness (Hamari & Koivisto, 2015; Ryan & Deci, 2000). For athletes, this is of particular importance stemming primarily from a highly passionate attitude towards their sport, resulting in a higher need for recognition for their achievements (Whelan

& Clohessy, 2021). Multiple studies have analyzed peer recognition in the context of QS applications: Hamari & Koivisto (2015) have studied perceived recognition in the context of an exercise gamification service and have found that recognition is positively correlated with usage continuation intention of the application, while Pinkerton et al. (2017) have found that

‘social media users are more likely to share details of their workout when they received peer recognition’ (Whelan & Clohessy, 2021; see also Pinkterton et al., 2017).

Thus, not only does peer recognition motivate users to continue using a fitness application in the future, the intensity of their in-app involvement increases simultaneously.

When individuals receive positive recognition from their peers, satisfaction for the person receiving recognition may be enhanced (Hamari & Koivisto, 2015; Kelman, 1958; Lin, 2008).

Furthermore, as found by Hamari & Koivisto (2015) and confirmed by several other researchers, the concept of recognition is closely tied to social acceptance, and may encourage humans to continuously conform to the anticipations of their digital health and fitness community over time, hereby potentially ensuring long-term usage continuation of the QS application in question (Kelman, 1961; Whelan & Clohessy, 2021). This is largely based on behavioral theories of operant behavior (Skinner, 1963) and social learning (Bandura, 1977), which emphasize that actions of positive reinforcement motivate individuals to continue

(18)

H1 There is a positive relationship between recognition and QS app usage continuation intention.

2.1.2 Reciprocal benefits

The second variable, reciprocal benefits, is a consequence of interpersonal interactions resulting in mutual advantages for users (Hamari & Koivisto, 2015; Hamari et al., 2014). In its essence, reciprocal benefits ‘may arise from group interactions, when receiving recognition is reciprocated by providing recognition’ (Hamari & Koivisto, 2015), thus creating an interconnected system of mutual positive reinforcement (Cialdini & Goldstein, 2004; Cialdini et al., 1992), as opposed to a one-sided positive feedback stream, which is characteristic of the recognition variable. Oftentimes, humans feel the social responsibility to reciprocate positive feedback provided by others, prompting them to return the favor as a method of signaling gratitude and acceptance towards other community members (Cialdini, 2009; Hamari &

Koivisto, 2015; Whelan & Clohessy, 2021). As found by Chiu et al. (2006), reciprocal interactions have the ability to enhance perceived usefulness of a social community system in individuals; hence, perceived mutual benefits created as a result of system use potentially increase when reciprocal interactions are occuring.

With regards to QS applications, reciprocal benefits have been classified as a main determinant of habitual app usage (Stragier et al., 2016). Hence, health and fitness applications are actively encouraging their user base to engage in reciprocal interactions; e.g., running app Strava displays the following message in their online environment: ‘You get more kudos when you give more kudos. Get out there and be a part of the conversation. Comment on friends’

activities and follow people you race with and pretty soon you’ll get more kudos than ever’

(Whelan & Clohessy, 2021; see also McCall, 2016). Besides, reciprocal benefits are a significant predictor of adherence to physical activity (Cavallo et al., 2012; Giles-Corti &

(19)

Donovan, 2002). Hence, as reciprocal benefits have been prominent in several previous research work conducted on the topic of social health and fitness applications, the following assumption is made:

H2 There is a positive relationship between perceived reciprocal benefits and QS app usage continuation intention.

2.1.3 Network exposure

Network exposure is the third variable of the social dimension and is entirely centered on the scope of one’s social network within QS applications (Hamari & Koivisto, 2015). Network exposure essentially refer to the size of one’s network and the hereupon based value that an individual perceives their network to have (Hamari & Koivisto, 2015; Hsieh et al., 2008; Lin

& Bhattacherjee, 2009). As defined by Hamari & Koivisto (2015), network exposure arises when ‘the benefits from using the service depend on the number of relevant other users’

(Hamari & Koivisto, 2015); thus, network exposure is more prominent the larger the user’s network. This dimension has been found to be imperative for services which are based on social interaction (Baker & White, 2010; Hamari & Koivisto, 2015; Lin & Lu, 2011; Sledgianowski

& Kulviwat, 2009), while additionally being a central element of social influence, as the size of one’s network is ‘proportional to the amount of influence a person can be exposed to’

(Hamari & Koivisto, 2015). The hypothesis is as follows:

H3 There is a positive relationship between network exposure and QS app usage continuation intention.

(20)

2.2 The moderating role of user personality type

To adequately answer the main research question, the personality traits extraversion, agreeableness, conscientiousness, neuroticism, and openness to experience are introduced as moderating variables. These variables take on a significant role within the context of mobile app adoption, technology usage, attitudes towards social media, and QS application usage (Alshawmar et al., 2022; Pinkerton et al., 2017; Xu et al., 2016).

Within the domain of personality psychology, vast consensus has been reached on using the five-factor model of personality as the primary model categorizing personality types of individuals (Digman, 1990; Guadagno et al., 2008; Halko & Kientz, 2010; John & Srivastava, 1999; McCrae & Costa, 1997). This model, often referred to as the Big Five, comprises of five dimensions, i.e., extraversion, agreeableness, conscientiousness, neuroticism, and openness to experience, and is ‘widely known as one of the major means of organizing human personality’

(Halko & Kientz, 2010).

Highly extraverted individuals are generally sociable, outgoing, and fun-loving, and tend to see the world as a place for constant social interaction. They have a strong need to share their thoughts and ideas with others, and to be engaged in permanent interaction with their social circle. The second personality dimension, agreeableness, is characterized by individuals who are kind by nature, constantly putting others’ needs and concerns ahead of theirs, and who behave self-sacrificing, compassionate, and polite towards others. They tend to stay away from conflict, as it is in direct opposition to their caring tendency towards other people. Individuals who are high in conscientiousness share character traits of being well-organized, prepared, self- disciplined, diligent, and orderly, almost to the extent of being industrious, which involves sacrificing the present in consideration of future returns. As the fourth personality dimension, neurotic individuals generally perceive their surrounding as dangerous, and have a tendency of being emotionally unstable. Furthermore, they tend to be easily stressed, self-pitying, and

(21)

anxious. Lastly, individuals with a high openness to experience find inspiration in their surroundings and are excited and interested to explore and learn new things. Due to their creative, imaginative, and variety-seeking nature, high openness to experience is directly correlated with intelligence (Guadagno et al., 2008).

With personality influencing virtually every area of a human’s life, differences in personality traits have the power to affect human interaction, thought patterns, and behavioral responses. Moreover, personality traits exert influence on an individual’s perception of technological innovations, social networking sites, communication technologies, and mobile applications. According to a study conducted by Unal et al. (2017), personality traits of an individual are primary determinants of their social media usage and general mobile phone use, whereas it has been previously established that personality differences are knowingly contributing to variations in Internet usage behavior of humans (Amichai-Hamburger & Ben- Artzi, 2003; Guadagno et al., 2008). Hence, the Big Five personality traits have been deemed as moderators in the context of this thesis, potentially affecting the relationship between how an individual perceives the social dimension of QS applications contingent on their personality type and its influence on an eventual app usage continuation intention. In the following section, main findings from prominent studies on the topic of personality types in relation to the topic of this thesis are presented, grouped by the five personality dimensions.

2.2.1 Extraversion

With regards to the first personality dimension, i.e., extraversion, two primary models have been developed to explain Internet usage behavior and attitudes towards online social networks of individuals scoring differently on the extraversion scale. According to the social enhancement hypothesis developed by Kraut et al. (2002), also called the ‘rich-get-richer’

(22)

many social connections in a real-world context. Contrarily, the social compensation hypothesis proposes that ‘the Internet will be used especially by introvert and socially anxious adolescents, who have difficulty developing friendships in their real-life environment’

(Valkenburg et al., 2005). Studies have yielded inconsistent results with regards to these models (Gross et al., 2002; Kraut et al., 2002; Zywica & Danowski, 2008); yet, several studies confirm that extraversion is correlated with larger social networks not only in the offline environment, but also in the context of social networking sites (Ong et al., 2011; Pollet et al., 2011; Swickert et al., 2002; Tong et al., 2008).

Furthermore, in a large-scale field study conducted by Xu et al. (2016), it has been found that a high level of extraversion is associated with mobile social app adoption. Not only are extraverts more likely to initially download a social app, extraversion is also positively correlated with usage continuation of social media, as extraverts display a significantly higher motivation for a continued, long-term usage of social networking sites due to their higher need to socialize with others (Alshawmar et al., 2022; Correa et al., 2010; Deng et al., 2013).

Furthermore, individuals scoring high on extraversion show a generally higher willingness to use social elements of Internet-based technologies (Lane & Manner, 2012; Rupp et al., 2018) and to create and share user-generated content, as Pinkerton et al. (2017) confirm that extraversion positively correlates with people’s willingness to share content about physical activities on social networking sites.

This high drive towards social involvement prompts extraverts to be continuously profoundly involved in social activities, both online and offline (Ross et al., 2009). Extraverts have been reportedly linked to having more group memberships in online communities (Ross et al., 2009), confirming that people with this personality trait strive for the maximum level of interpersonal interaction both in the real-life context as well as in the digital environment. This is largely due to extraverts being ‘more likely to seek opportunities for stimulation and

(23)

interaction that exist outside of themselves’ (Kandler et al., 2014; see also Rupp et al., 2018), which is highly distinctive for this personality trait. Besides, studies confirm that extraversion improves happiness as a consequence of receiving social support, which in turn is partially dependent on network size in the digital setting (Tan et al., 2018). Based on the existing literature, it is expected that extraverts’ high tendency to socialize also applies in a QS application context; thus, the following hypothesis is formulated:

H3a The relationship between network exposure and QS app usage continuation intention is positively moderated by extraversion.

2.2.2 Agreeableness

As the second personality dimension, agreeableness is a significant predictor of high trust towards technology, which further translates to a high trust towards social networking sites (Digman, 1990; Rupp et al., 2018). Highly agreeable individuals show a greater frequency of social media usage, with their reported reasons for using social networking sites being 1) to obtain informational content, including news and 2) to interact with others (Gil de Zúñiga et al., 2017). Studies have found that agreeableness is significantly correlated with ‘needs for belonging (connection and caring), communication uses of social media, and self-presentation online’ (Gil de Zúñiga et al., 2017; see also Seidman, 2013), while agreeable individuals show a high likelihood of using social media to maintain connections with others (Seidman, 2013).

As agreeable individuals are highly people-oriented, belongingness concerns on social media are highly relevant for people with this personality trait (Seidman, 2013). Furthermore, evidence has shown that highly agreeable people tend to have a greater support satisfaction within their digital social network, indicating that agreeable individuals are more likely to feel satisfied with the social support they receive in the digital context (Pornsakulvanich, 2017).

(24)

It has been found that agreeableness is the ‘strongest predictor of prosocial behavior among the Big Five’ (Byerly et al., 2022) personality traits, signifying that agreeable individuals are not only profoundly oriented towards others, but that they also have a tendency to prioritize the needs of other people (Habashi et al., 2016). This holds true for real-life interactions as well as for online interactions, as Leng et al. (2020) confirm that agreeableness is positively correlated with exhibiting prosocial behavior online. Therefore, the following hypothesis is being tested:

H2a The relationship between perceived reciprocal benefits and QS app usage continuation intention is positively moderated by agreeableness.

2.2.3 Conscientiousness

Thirdly, regarding the personality type of conscientiousness, research has found that high levels of this dimension predict satisfaction with socially-based technologies, simultaneously increasing highly conscientious people’s willingness for using these technologies (Halko &

Kientz, 2010). These findings are in line with the study by Alshawmar et al. (2022), which was conducted in the context of health and fitness applications, and which led to the conclusion that highly conscientious individuals perceive a higher usefulness of QS applications, whereas they particularly value competition and cooperation elements (Chen & Pu, 2014; Halko & Kientz, 2010). Although it has been acknowledged in several studies that high levels of conscientiousness inhibit group performance due to conscientious people excelling at performing structured tasks and working diligently individually rather than in a group setting (Hakim et al., 2021), the study by Halko & Kientz (2010) conducted within the context of QS applications concluded that high levels of conscientiousness significantly correlate with willingness to engage in cooperative strategies to complete fitness goals together. These findings can be explained by a high goal-orientation of conscientious people, which appears to

(25)

not only apply in a self-centered sense, but which also carries over to group interactions (Halko

& Kientz, 2010). Thus, the following hypothesis is tested:

H2b The relationship between perceived reciprocal benefits and QS app usage continuation intention is positively moderated by conscientiousness.

2.2.4 Neuroticism

With regards to neuroticism, the study conducted by Xu et al. (2016) has shown that neuroticism is positively correlated with mobile social app adoption, indicating that individuals scoring high on this dimension have a strong tendency for downloading mobile apps, which incorporate social interaction features. Besides, neurotic females show a higher need for interacting with their social network in a digital setting (Guadagno et al., 2008; Hamburger &

Ben-Artzi, 2000), making the ‘inclusion of […] socializer design features […] a high priority’

(Anvari et al., 2017) to meet the need of neurotic individuals in a digital application context. It has been generally established that neurotic individuals feel more comfortable expressing themselves in an online setting as opposed to an offline setting (Amichai-Hamburger et al., 2002; Guadagno et al., 2008), while neuroticism is generally correlated with a preference for indirect and non-confrontational communication forms (Butt & Phillips, 2008). Due to their anxious nature, neurotic people have a higher preference for communicating via social media sites and other types of written communication as opposed to direct interactions involving calls and face-to-face conversations (Butt & Phillips, 2008). Taking these findings into account, it can be hypothesized that having a large network in the digital environment is important for neurotic individuals, as it allows people scoring high on this personality dimension to socially interact with a large number of people in a non-confrontational manner. The following hypothesis is formulated:

(26)

H3b The relationship between network exposure and QS app usage continuation intention is positively moderated by neuroticism.

2.2.5 Openness to experience

As the final personality dimension of the Big Five, openness to experience is directly correlated with social media use (Correa et al., 2010), where a high level of openness predicts usage of social networking pages for the purpose of gaining novel experiences (Ross et al., 2009). The same study confirms that highly open individuals are more likely to behave sociable on social networks (Ross et al., 2009), which is in line with their general curiosity towards new situations and other human beings. Not only are people who have a high openness to experience more likely to be outgoing on social media, they are also more likely to be more self-disclosing, as studies confirm that people with a high openness ‘share more personal information on Facebook profiles’ (Hong & Oh, 2020; see also Amichai-Hamburger & Vinitzky, 2010;

Guadagno et al., 2008). Furthermore, people having a high openness to experience are more likely to blog on digital channels, in which they are fully open to share details about their personal life with the online community (Guadagno et al., 2008). Due to the parallels of blogging with the recognition dimension, it can be hypothesized that receiving peer recognition for sharing content on QS applications leads to a greater app usage continuation intention for individuals with a high openness to experience. The hypothesis is as follows:

H1a The relationship between recognition and QS app usage continuation intention is positively moderated by openness to experience.

2.3 Effect of gender

As the control variable of this study, gender of application users is considered an influencing factor in the relationship between the social dimension and QS app usage continuation

(27)

intention. Not only have studies confirmed significant gender differences in motivation to participate in physical activities (Molanorouzi et al., 2015), substantial dissimilarities also exist pertaining to technology use, particularly health and fitness application usage, and judgements of the social dimension within online networks (Mouakket, 2018). Regarding technology use, studies have found that males and females differ significantly not only in the degree to which they favor technology, but also in the purpose of their usage. Overall, men tend to have a more positive attitude towards technology compared to women (Cai et al., 2017), which has been confirmed in several studies (Ardies et al., 2015; Ong & Lai, 2006; Sáinz & López-Sáez, 2010;

Yau & Cheng, 2012). Yet, while the attitude of females towards technology use is generally lower compared to the attitude of males, it still falls within the positive spectrum, indicating that both men and women have a generally favorable attitude towards technology, although at a higher degree for males (Cai et al., 2017).

The intended purpose of technology usage differs by gender: While males use Internet- based technologies predominantly in a task-oriented manner, e.g., for reading the news or acquiring factual information, females tend to view technology as a means for relationship- building tasks, social interaction, and interpersonal communication (Kimbrough et al., 2013;

Weiser, 2000). This finding is in line with a generally high relationship orientation of females, which pervades many areas of technology for women. Research conducted by Kimbrough et al. (2013) confirms that females show a preference for the social elements of technology, including social media and textual communication with their network, resulting in the finding that females connect more with each other in the digital sphere compared to males.

Pertaining to mobile health and fitness applications, this tendency continues: Whereas for women the most vital antecedents for QS app adoption are ‘well-being, stress, and social elements (i.e., friendships)’ (Antezana et al., 2022), men indicated that the functionality of a

(28)

adoption on their part (Antezana et al., 2022). Furthermore, satisfaction with a social, digital platform plays a more central role for males compared to females (Mouakket, 2018), reconfirming the notion that functionality and satisfaction hereof are crucial for males not only for social app adoption, but also for determining continuance intention.

Furthermore, prominent gender differences pertaining to the participation in physical activity have been confirmed in previous literature. While females derive their motivation to participate in physical exercise predominantly from extrinsic factors, such as physical appearance, attractiveness, and weight management (Egli et al., 2011; Molanorouzi et al., 2015), males are primarily motivated by factors relating to competition, challenge, and strength, hereby driven by the overarching intrinsic goal of achieving mastery in their sport (Egli et al., 2011; Molanorouzi et al., 2015; Morgan et al., 2003; Morris et al., 1995). Thus, it is analyzed whether the elements of the social dimension, including the competition-related recognition dimension, have a different impact on males as compared to females.

(29)

3. Conceptual framework

Figure 1

Conceptual Framework

The conceptual framework considers the social dimension as a central element impacting the relationship with QS app usage continuation intention. In line with Hamari & Koivisto (2015) and Whelan & Clohessy (2021), the social dimension is hereby categorized into three fundamental variables, namely recognition, reciprocal benefits, and network exposure, acting as the independent variables (IVs) in this framework. QS app usage continuation intention is the dependent variable (DV) of this study, subject to the influence of the three IVs. Gender is acting as a control variable, whereas the five personality types are the moderating variables.

Main effects

H1 There is a positive relationship between recognition and QS app usage continuation

Extraversion

QS app usage continuation intention Recognition

Reciprocal benefits

Network exposure

Gender Conscientiousness

Neuroticism Agreeableness

Openness to experience

H1

H2

H3

H1a H2a

H2b

H3a H3b

(30)

H2 There is a positive relationship between perceived reciprocal benefits and QS app usage continuation intention.

H3 There is a positive relationship between network exposure and QS app usage continuation intention.

Interaction effect: Recognition

H1a The relationship between recognition and QS app usage continuation intention is positively moderated by openness to experience.

Interaction effect: Reciprocal benefits

H2a The relationship between perceived reciprocal benefits and QS app usage continuation intention is positively moderated by agreeableness.

H2b The relationship between perceived reciprocal benefits and QS app usage continuation intention is positively moderated by conscientiousness.

Interaction effect: Network exposure

H3a The relationship between network exposure and QS app usage continuation intention is positively moderated by extraversion.

H3b The relationship between network exposure and QS app usage continuation intention is positively moderated by neuroticism.

(31)

4. Research method

4.1 Research design

To scientifically test the hypotheses, primary quantitative data was collected using the digital data collection tool Qualtrics. An electronic survey has been created and distributed to the sample population, testing the effect of the independent variables (IVs) recognition, reciprocal benefits, and network exposure on the dependent variable (DV) QS app usage continuation intention, under the moderating effect of extraversion, agreeableness, conscientiousness, neuroticism, and openness to experience. Measuring actual QS app usage continuation would have required an extensive experimental design involving the development of an application for this particular purpose; hence, it has been decided to measure continuation intention by means of a quantitative study.

Part of the survey has been framed as a scenario study, instructing participants to imagine themselves in a hypothetical situation of using a fitness application and browsing their feed in the app environment. As convenience sampling was the chosen sampling method of this study, the survey has been distributed in the personal network of the author; next to this, the digital data collection platforms Surveyswap and Surveycircle were used. The time frame of data collection was nine days, and took place from May 22, 2022 to May 31, 2022.

4.2 Description of sample

The sampling technique applied in this study was convenience sampling, as part of non- probability sampling. As opposed to restricting the sample population to past and current users of health and fitness applications, all user types (including nonusers) were eligible to participate in the study. Participants indicated their user type in the beginning of the survey. There were

(32)

average statistics, the estimated response rate was 33%. In total, 230 responses were collected during the time frame of data collection.

4.3 Procedure and description of measures

All measures were self-reported, as respondents were asked to provide information on their own behavior, outlooks, and intentions; hence, all measures were obtained from one source.

Furthermore, all measures were obtained at the same time point, i.e., within the time frame of the data collection. Validated measures were available for all variables in the model; besides, all variables were assessed in a closed-end format.

In the first section of the survey, respondents were informed about the purpose and procedure of the study; furthermore, information was provided about anonymity and the possibility to withdraw from the study at any given time. Consent was obtained from respondents before the start of the survey. Following the initial information, respondents were asked to provide demographic information, including their gender, age, country of residence, and highest level of education completed. Gender and level of education were list questions, whereas age and country of residence were open questions. Hereafter, to assess users’

personality type, the measurement by Donnellan et al. (2006) has been incorporated as a shortened version of the Big-Five personality type questionnaire. This 20-item questionnaire determines the 5 personality dimensions based on a 7-point Likert scale (1 = strongly disagree, 7 = strongly agree). An example of a question is ‘I am not interested in other people’s problems’, indicating low levels of agreeableness.

Hereafter, a list question was asked to determine respondents’ frequency of physical activity participation (e.g., 3-4 times per week). Subsequently, respondents were asked to specify their usage status regarding fitness apps in a list question, to determine whether they are current users, past users, or nonusers of QS applications. Two conditional questions were

(33)

following for respondents who are currently using one or multiple health and fitness app(s), namely a list question addressing the frequency of use (e.g., weekly), and the usage duration (e.g., a few weeks). These measures were based on a study by Alshawmar et al. (2022).

Hereafter, a scenario has been introduced to the respondents. Participants were asked to imagine that they have downloaded a fitness app which incorporates social features. They were asked to imagine that they are scrolling through the activity feed on the app, in which they see their networks’ posts, physical accomplishments, and an invitation to join a challenge.

Following this scenario, a measure by Hamari & Koivisto (2015) was included to address respondents’ views on the social dimension, further subdivided into recognition, reciprocal benefits, and network exposure. This measure included 12 items on a 7-point Likert scale (1 = strongly disagree, 7 = strongly agree), with an example question being ‘I feel good when my achievements in my fitness app are noticed’, indicating a high level of recognition.

As Hamari & Koivisto’s (2015) study was conducted within the boundaries of a single QS application, i.e., Fitocracy, the questionnaire of the current study has been slightly adapted to include fitness applications in a broader sense, replacing ‘Fitocracy’ with ‘my fitness app’.

Besides, as Hamari & Koivisto (2015) measured the perception of current users, the items measuring network exposure required minor adaptations, as the present sample does not necessarily have an existent QS application network of friends, family members, and other athletes depending on their user status. Thus, items were modified to reflect a hypothetical scenario, with an example statement of Hamari & Koivisto’s (2015) study being ‘I have many friends in Fitocracy’, adapted to 'Having many friends on my fitness app is important to me’.

Lastly, to measure app usage continuation intention, a shortened, 2-item measure by Hamari & Koivisto (2015) has been incorporated, with an example item being ‘I predict that I will keep using this app in the future as much as I have used it lately’, indicating a high app

(34)

strongly disagree, 7 = strongly agree). The complete survey questionnaire can be found in Appendix 1.

(35)

5. Data analysis and results

5.1 Respondents

A total of 230 responses were collected during the time frame of data collection through the online tool Qualtrics. Included in this number were partial responses of participants, who dropped out throughout the course of the survey, or who left questions partially unanswered.

193 out of 230 respondents filled out the survey with a completion rate of 100%, leading to a response rate of 83.9%. 37 participants did not complete the survey. Furthermore, 11 respondents out of the included 193 responses failed the attention check integrated towards the end of the survey, leading to a total of 182 responses incorporated in a further analysis (further information regarding data cleaning can be found in chapter 5.2.1).

Out of 182 respondents, 63.2% were female, 34.6% were male, and 2.2% identified as non-binary. Regarding the geographic distribution, the largest group of respondents (46.7%) identified the Netherlands as their current country of residence, followed by the United Kingdom (12.1%), Germany (8.2%), and the United States (8.2%). Respondents indicated an age ranging from 18 to 68, with a mean of 26 and a standard deviation of 7.3. The majority of respondents appeared to be highly educated, with 81.8% holding a Bachelor’s, Master’s, or PhD degree. An in-depth overview relating to demographic information of the respondents can be found in Appendix 2.

With regards to the current level of physical activity participation and respondents’ QS app usage, the following findings were made: The majority of respondents shows moderate levels of physical activity participation, with 46.2% indicating to exercise 1-2 times per week, and 26.4% indicating to exercise 3-4 times per week. 17% of respondents exhibit high levels of physical activity participation, with 7.1% being active daily and 9.9% exercising 5-6 times per week on average. 10.4% of respondents indicated that they never exercise. Based on 182

(36)

41.8% claim to have used a health and fitness application in the past, yet have stopped using it. 23.1% have never used a health and fitness app and therefore classify as nonusers.

Respondents who are current users of QS applications have been asked two conditional follow- up questions to determine their app usage frequency and duration. Based on 64 current users, 40.6% claim to use their QS application(s) daily, while 50% of respondents use their QS app(s) weekly. The minority (9.4%) indicates to use their QS app(s) on a monthly basis. Furthermore, the majority of current users has been using the QS app long-term, with 56.3% claiming to have used the application(s) for a few years. 32.8% of respondents have used their health and fitness application(s) for a few months, while 10.9% have used it for a few weeks. A comprehensive overview regarding physical activity and QS app usage of respondents can be found in Appendix 3.

5.2 Data preparation

5.2.1 Data cleaning and dealing with missing values

All variables of the conceptual model were analyzed for missing data by running a frequency test in SPSS. Missing data took up a percentage of <10% for all variables except for recognition, reciprocal benefits, network exposure, and QS app usage continuation intention.

Hereafter, missing values were omitted from a further analysis by excluding cases listwise. Out of a total of 230 responses, 193 responses were left after excluding all cases with missing values. Ultimately, 182 responses were included in a further analysis after filtering out responses which failed the attention check during the questionnaire.

5.2.2 Recoding variables

As gender is a moderating variable in this study, the gender scale has been adapted from 1 (Male), 2 (Female) to 0 (Male), 1 (Female) to prepare the data for the analysis. In the next step,

(37)

counter-indicative items have been recoded; the only variable where this was applicable is user personality type. This means that items which, if participants selected a high number on the 7- point Likert scale, refer to a low level of the measured construct. The items rPersonality6, rPersonality7, rPersonality8, rPersonality9, rPersonality10, rPersonality15, rPersonality16, rPersonality17, rPersonality18, rPersonality19, and rPersonality20 were recoded and reversed;

thus, the items RePersonality6, RePersonality7, RePersonality8, RePersonality9, RePersonality10, RePersonality15, RePersonality16, RePersonality17, RePersonality18, RePersonality19, and RePersonality20 were created.

5.2.3 Computing reliability

To determine the internal consistency of the measurement scales, a reliability test was performed to assess the Cronbach’s alpha for each scale. For the variables recognition, reciprocal benefits, network exposure, extraversion, agreeableness, conscientiousness, and continuation intention, a Cronbach’s alpha value of > .7 was found, confirming the internal consistency of their scales (Field, 2018). For the variables neuroticism and openness to experience, a value of < .7 was detected (neuroticism: =.642; openness to experience:

=.695). Regarding neuroticism, omitting the item ReNeuro19 referring to the statement ‘I seldom feel blue (=sad)’ from the dataset would increase the Cronbach’s alpha score to > .7.

Yet, this item was retained since omitting it would result in the number of items measuring this personality trait decreasing from four to three, which would threaten the construct validity of this measure. For the variable openness to experience, no items were omitted as an exclusion of statements would not have resulted in an increase of the Cronbach’s alpha score to > .7.

(38)

Table 1

Cronbach’s Alpha

Variable Cronbach’s Alpha

Recognition .920

Reciprocal benefits .923

Network exposure .929

Extraversion .803

Agreeableness .781

Conscientiousness .720

Neuroticism .642

Openness to experience .695

Continuation intention .802

5.2.4 Normality, linearity, and homoscedasticity

A normality check was undertaken to determine the distribution of residuals. The two items measuring QS app usage continuation intention of the respondents (ContInt1, ContInt2) were combined and the item ContInt was created. Hereafter, histograms and P-P plots were created for all three IVs in relation to the DV (see Appendix 4), indicating an approximate normal distribution for the variables in this model.

To assess the linearity between the IVs and the DV, multiple scatterplots have been created (see Appendix 5), confirming a positive linear associations between all three IVs and the DV. Lastly, a scatter plot has been created to assess homoscedasticity of residuals (see Appendix 6), which indicates that the residuals are equally constant; hence, the data is homoscedastic. Therefore, all assumptions of the regression are met.

(39)

5.3 Correlational analysis Table 2

Means, Standard Deviations, and Correlations of Variables

Variable M SD 1 2 3 4 5 6 7 8 9 10

1. Recognition 4.52 1.501 -

2. Reciprocal benefits 4.67 1.494 .764** - 3. Network exposure 3.13 1.456 .656** .583** -

4. Extraversion 3.97 1.306 .190* .096 .109 -

5. Agreeableness 5.36 .997 .133 .146* .005 .352** -

6. Conscientiousness 4.73 1.116 -.031 .003 .037 .024 .017 -

7. Neuroticism 3.99 1.044 .019 -.016 .032 -.071 .100 -.270** -

8. Openness to experience 5.18 .960 .074 .066 -.109 .116 .303** -.145 .095 -

9. QS app usage continuation intention

3.99 1.486 .385** .335** .425** .176* .079 .085 -.003 .069 -

10. Gender¹ .68 .514 -.024 -.027 -.020 .024 .076 .008 .124 -.057 .011 -

Note. N = 182. ¹Gender was coded as 0 = male and 1 = female.

*p < .05 ** p < .01 (two-tailed)

The results of the correlational analysis (Table 2) indicate that recognition has the highest standard deviation in this dataset (SDRecognition=1.501, SDReciprocal benefits=1.494, SDNetwork exposure=1.456, SDExtraversion=1.306, SDAgreeableness=.997, SDConscientiousness=1.116, SDNeuroticism=1.044, SDOpenness to experience=.960, SDQS app usage continuation intention=1.486). This implies that it has the largest variance compared to the other variables in the model. Agreeableness has the highest mean score (M=5.36), indicating a tendency of respondents towards ‘strongly agree’, whereas network exposure has the lowest mean score (M=3.13) implying that respondents were more inclined to answer on the left side of the 7-point Likert scale, towards

‘strongly disagree’.

There are positive correlations between the independent variables, as recognition is positively correlated with reciprocal benefits (r=.764, p<.001), as well as with network exposure (r=.656, p<.001). Furthermore, network exposure is positively correlated with

(40)

independent variables tend to score high on all independent variables, whereas respondents scoring low on one IV tend to score low on the other two IVs simultaneously. With regards to the moderating personality types, it can be concluded that there is a significant positive relationship between extraversion and agreeableness (r=.352, p<.001), agreeableness and openness to experience (r=.303, p<.001), as well as a significant negative relationship between conscientiousness and neuroticism (r=-.270, p<.001). This implies that based on this dataset, extraverted individuals tend to be highly agreeable, whereas agreeable individuals tend to simultaneously score high on openness to experience. The negative relationship between conscientiousness and neuroticism implies that respondents scoring high on conscientiousness tend to score low on neuroticism and vice versa. Moreover, there is a marginal negative trend between conscientiousness and openness to experience (r=-.145, p<.100).

Significant relationships were also found between the moderators and the independent variables in this model. Recognition shows a significant positive correlation with extraversion (r=.190, p=0.10), which entails that extraverted individuals score high on the recognition scale, whereas introverted respondents score low. Furthermore, the IV reciprocal benefits is positively correlated with agreeableness (r=.146, p<.050), indicating a relationship between those two variables. Lastly, there is a marginal positive trend between recognition and agreeableness (r=.133, p<.100).

Pertaining to the dependent variable QS app usage continuation intention, several significant relationships have been found: Not only is the DV positively correlated with all three moderators of this model (recognition: r=.385, p<.001; reciprocal benefits: r=.335, p<.001; network exposure: r=.425, p<.001), there is also a significant positive relationship between extraversion and QS app usage continuation intention (r=.176, p<.050). The control variable gender does not show a significant relationship with any of the variables; the only variable with a marginal positive trend is neuroticism (r=.124, p<.100), indicating that there is

Master’s Thesis

Master’s Thesis

Statement of originality

Table of Contents

List of tables and figures

Acknowledgements

Abstract

1. Introduction

2. Literature review

3. Conceptual framework

4. Research method

5. Data analysis and results