The hypotheses

The model and related hypotheses are based on the functionalities and features that are an integral part of the self-tracking device’s design. The study focuses on investigating how the functionalities of the device affect the participant’s personal daily performance in relation to the steps and sleep. The data surrounding these functionalities are tracked and collected entirely through the Jawbone UP app and the RescueTime app. Therefore, the model is specifically tailored to the Jawbone UP device and its functionalities as a self-tracking device.

This study involves two linear regressions, each related to steps and sleep, respectively. They are different activities with different goals set into the device.

This means that step performance might be satisfactory whereas sleep performance is not. Moreover, the different activities also differ in how the participant is able to influence the performance outcome. For example, the participant can actively choose to take a walk around the block to get more steps, whereas it is more difficult (though not impossible) for the participant to actively choose to fall back asleep for another hour to improve the sleeping statistics.

The hypotheses of the linear regressions are grouped into three categories:

behavior, engagement and social elements. These categories are based on the functionalities stemming from the self-tracking device. As an outset, the behavior category concerns the participant’s performance that is captured by the device and then showcased as personal data in the app. It is thus actual behavioral measures that are embodied by the technology, as described by Yoo (2010). The behavior category is addressed first because it is the behavior that achieves a high success in relation to the performance goal and therefore, it is likely that the behavior has a great influence.

The engagement category is related to the user’s participation and interaction with the mobile app and data. The role of engagement is an important part of this study because it has been identified in the theoretical background. More specifically, engagement with personal data is identified to support the process of self-reflection, which in turn has a positive influence on behavior, such as performance measures like steps and sleep (e.g., Consolvo et al., 2009; Li, Dey, & Forlizzi, 2012; Lin et al., 2006). Although engagement can take several forms, most importantly it includes the exposure of data to the participant. In this study, engagement is thus understood as participation with the personal data offered by the app that is related to the self-tracking device (Choe, Lee, Lee, Pratt, & Kientz, 2014; Sjöklint, Constantiou, & Trier, 2015). It is thus understood as having both interaction with and observation of the data.

The social category involves the social aspects available through the device, such as the user inviting other users to see the data and in turn being exposed to their data. The social element is important to discuss, for it is argued that it has an influence on the participant, who feels pressure and need to conform to the behavior and expectations of others, which could influence performance (Aaronson et al., 1994).

As an overview, the following are the hypotheses relating to step performance:

H1: Being active has a positive influence on the step performance.

H4: Checking the app has a positive influence on step performance.

H6: Increasing the step goal is negatively associated with the step performance.

H8: Notifications are positively associated with step performance.

H10: Social connections in the mobile app have a positive influence on the step performance.

The following are the hypotheses relating to sleep performance:

H2: The amount of deep sleep has a positive influence on sleep performance.

H3: Going to bed before 23.00 has a positive influence on sleep performance.

H5: Checking the app has a positive influence on sleep performance.

H7: Increasing the sleep goal is negatively associated with the sleep performance.

H9: Notifications are positively associated with sleep performance.

H11: Social connections in the mobile app have a positive influence on the sleep performance.

5.2.1 Behavior

The first category involves the behavior of the user, which is captured and embodied by the self-tracking device. When it comes to steps, the Jawbone UP records the steps, but also when the steps stem from an activity, such as taking a run, or those that are in an idle setting. As for sleep, the Jawbone UP records sleep phases, such as light sleep versus deep sleep and whether the user woke up during the night.

5.2.1.1 Step related behavior

The second category is behavior and includes user activity. The user’s lifestyle in relation to the level of activity is likely to have an impact on the step performance.

For example, adults who have a history of being physically active in their youth are 2-3 times more likely to be active as they age (Dishman, Sallis, & Orenstein, 1985). Consistent activity is thus an important part of staying active. This is also reflected in the use of an activity tracker, such as a pedometer, since the users who are initially leading an active lifestyle are more likely to perform better compared to those who have led more sedentary lifestyles (Tudor-Locke et al., 2004).

However, it has also been shown that those who were not physically active but had a desire to change also increased activity that was sustainable over time (J. J. Lin et al., 2006). Overall, user activity is associated with both intrinsic and extrinsic motivation, which means that the user may already be active before wearing a self-tracking device, yet still gain motivation from wearing it (Li, 1999). The hypothesis is as follows:

H1: Being active has a positive influence on the step performance.

5.2.1.2 Sleep related behavior

Deep sleep is one of the indicators of the quality of the user’s sleep session. While asleep, we go through several phases, which can be generally referred to as light sleep and deep sleep. Light sleep is usually 55% whereas deep sleep is 20% of the

sleep duration (K. A. Lee, Zaffke, & McEnany, 2000). The deep sleep is important for the person to feel more rested the next day (Horne, 1990). Deep sleep is thus an essential part of sleep and occurs in cycles, which means that more deep sleep generally means a longer sleep session (Babloyantz, 1986). The Jawbone UP app’s sleep functions automatically measure the phases of deep sleep and showcase this to the user upon request. The accuracy of the measurement may be debated (Rettner, 2014), yet this study focuses on what occurs during and after exposure of the data, so the importance is placed on the fact that the device is consistent in its capture and reporting to the user. A longer period of accumulated deep sleep is considered to be an important indication that the user has slept for a longer duration, and therefore performs better. Therefore, the hypothesis is:

H2: The amount of deep sleep has a positive influence on sleep performance.

The exposure to sleep data may cause the participant to gain awareness that leads to changing behavior so as to perform better on the sleep measurements. This is because engagement with the data can help individuals to “increase their awareness and encourage healthy behavior change” (Choe, Consolvo, Watson, &

Kientz, 2011, p.3060). This inspiration for behavioral changes may be going to bed early or at a consistent time, in favor of getting a better sleep experience (Stepanski & Wyatt, 2003). A study showed that going to bed at 23.40 was too late and associated with a decreased quality in the sleep experience (Buboltz, Brown, & Soper, 2001; Buboltz et al., 2009). Therefore, the hypothesis is:

H3: Going to bed before 23.00 has a positive influence on sleep performance.

5.2.2 Engagement

The second category is engagement and includes hypotheses regarding variables such as checking the app, changing goals and notifications. The data collected on these variables all stem from participation and interaction with Jawbone UP functionalities.

5.2.2.1 Checking the app

The first step of engagement with the personal data occurs by simply checking the Jawbone UP app. The Jawbone UP is understood as a commitment device that

helps the participant to act more deliberately and consciously—it reminds the user of the initial commitment (Ariely & Wertenbroch, 2002; Milkman et al., 2008). In this study, the Jawbone UP is thus a commitment device where the user commits to a more active lifestyle through steps and sleep. By wearing the device, the user is encouraged to act more deliberately by collecting and subsequently checking the personal data gathered by the mobile app. This is because checking the data leads to greater awareness of personal behavior and influences the user to do what should be done, rather than what he or she wants to do (ibid). The influence of different types of commitment devices has been studied in a self-tracking context, as well. For example, a study on the use of pedometers argues that wearing a self-tracker that shows the user personal data leads to activity increase (Chan, Ryan, &

Tudor-Locke, 2004; I. Li, 2012). Therefore, the hypothesis is:

H4: Checking the app has a positive influence on step performance.

Using the same assumption as above, sleep performance is also believed to be influenced by continuously checking the app. For example, one study showed that becoming more aware about sleep patterns led to sleeping better because the individual made changes to his or her lifestyle (Stepanski & Wyatt, 2003).

H5: Checking the app has a positive influence on sleep performance.

5.2.2.2 Changing the goal

Moreover, changing the goal can have alternative effects on the personal performance, depending on whether the performance is increased or decreased.

For example, this study asserts that changing the goal by increasing it has a negative impact on the personal performance because it becomes more difficult for the user to reach it. A study on step goal-related behavior showed that the user might be inspired to change the goal upward alongside recently increased activity.

At the same time, the user’s activities and practices may change in such a way that the goal is not met, even if the user is behaving more actively overall (Fritz et al., 2014). For example, if a user takes on yoga, the user is becoming more active, yet the added activity might not match the increased step goal. The pilot study showed that the users were hesitant to decrease the goal, but considered increasing it.

Based on these considerations, the hypothesis is:

H6: Increasing the step goal is negatively associated with the step performance.

On the basis of the discussion above about step goals, the same assumption is adopted for sleep-related goals. The hypothesis is thus:

H7: Increasing the sleep goal is negatively associated with the sleep performance.

5.2.2.3 Notifications

The Jawbone UP has the functionality of sending the user different types of notifications about steps and sleep. In a self-tracking context, notifications are argued to be a useful reminder to users to do a task (Fogg, 2009). Another study support this idea and extends it by endorsing the power of mobile notifications to situations where users are encouraged to manually log behavior and performance as well (Bentley & Tollmar, 2013). The notifications are helpful to get the user to do something by bringing greater awareness. As a result, in this study notifications are argued to be useful for bringing awareness to the user to do the tasks related to the Jawbone UP. Therefore, the hypothesis is:

H8: Notifications are positively associated with step performance.

On the basis of the same considerations, the assumption is applied to sleep performance as well. The hypothesis is:

H9: Notifications are positively associated with sleep performance.

5.2.3 Social elements

The third category incorporates the presence and possible influence of social elements on users. The social data is collected by viewing the user’s account and identifying how many other users have been invited to share the data. The social elements can be considered a type of engagement with others and their data.

The incorporation of social connections can contribute by adding motivation for the user to be more active as a response to social pressure. The presence of another

user may influence the self-tracker to strive towards being accepted, according to the norms existing in the setting (Aaronson et al., 1994). A study on social pressure showed that people were indeed more prone to comply with any formal or informal codes in the presence of others, yet also less inclined to comply when others did not (Wogalter, Allison, & McKenna, 1989). In a self-tracking context, collective use is shown to offer social support (Ploderer et al., 2014). Therefore, the hypothesis is:

H10: Social connections in the mobile app have a positive influence on the step performance.

Using the same background as social and steps performance, social influence has been argued to provide motivation for the user to conform to the contextual norms.

In a self-tracking context, social factors may also influence sleep performance, as in the user making some attempt to synchronize times with other people who are around (E. Choe et al., 2011). As sleep data is considered very private, it is important to find the right community to share it with (Fritz et al., 2014).

H11: Social connections in the mobile app have a positive influence on the sleep performance.