Self-tracking and gamification of physical activity: Effects on wellbeing Elaine Marie Grech 1, Marie Briguglio 1 and Emanuel Said 1 1 University of Malta, Msida, MSD2080, Malta Abstract The use of physical activity trackers and gamification can have wide-ranging benefits, however, their effect on wellbeing has not yet been sufficiently examined in literature. Our study examines whether gamified and non-gamified self-tracking experiences create positive psychological responses that yield enhanced wellbeing. We gathered data on self-reported happiness and life satisfaction before and after a four-week self-tracking experience of physical activity, with and without the use of gamification. We measured the users’ emotional and cognitive responses and parsed out the effect of these psychological outcomes on the users’ subjective wellbeing. We found that users’ perceived usefulness of the experience was associated with an increase in the individuals’ life satisfaction, while enjoyment and interest were linked to an increase in the individuals’ happiness. Both gamified and non-gamified self- tracking experiences evoked similar positive emotional and cognitive responses, yielding similar gains in wellbeing. Future research could explore long-term health and wellbeing impacts. Keywords 1 Self-tracking; gamification; psychological outcomes; life satisfaction; happiness; wellbeing 1. Introduction gather data on wellbeing measures (life satisfaction and happiness) before and after an experience of self-tracking alone, and in In the realm of physical activity, exercise and conjunction with the use of gamification. We wellbeing, there is widespread use of wearables, measure the users’ emotional (enjoyment and physical activity trackers, mobile fitness interest) and cognitive (perceived usefulness) applications, and extensive use of gamification to responses and parse out the effect of these promote and maintain regular physical activity psychological outcomes on the users’ subjective [1]. The aim of these technologies is arguably to wellbeing. We also examine whether gamification provide the motivational enforcement through enhances the effects compared to a non-gamified self-knowledge on one’s physical activity, goal self-tracking experience. setting, social influence, and social support to Through the findings of this study, we extend achieve self-improvement goals [2], [3]. our understanding of the psychological responses Literature suggests that self-tracking experiences that enhance wellbeing and contribute to the and gamification can have beneficial effects on literature on gamification and self-tracking. the users’ wellbeing [4], [5]. However, the effect of these behavioral interventions on wellbeing has not been sufficiently examined [4], [6]–[8]. 2. Literature We question whether experiences of self- 2.1. Theoretical underpinnings tracking and gamification create positive emotional and cognitive responses that yield Self-tracking technologies (also referred to as enhanced wellbeing. To answer this question, we quantified self) [9] enable people to collect data 7th International GamiFIN Conference 2023 (GamiFIN 2023), April 18-21, 2023, Lapland, Finland. EMAIL: elaine.m.grech@um.edu.mt (A.1) ORCID: 0000-0002-0856-0374 (A.1); 0000-0003-4018-795X (A.2); 0000-0002-6481-6258 (A.3) ©️ 2023 Copyright for this paper by its authors. Use permitted under Creative Commons License Attribution 4.0 International (CC BY 4.0). CEUR Workshop Proceedings (CEUR-WS.org) 119 about themselves. Physical activity trackers help religious participation, socialization, environment people realize their level of physical activity or quality and cultural participation [26]. rather inactivity. The provision of personal Motivational design technologies which informatics data received through the use of promote physical activity can also enhance activity trackers initiates a process of self- wellbeing outcomes. The premise is based on the reflection and evaluation [10], [11]. This process underlying processes that these interventions can brings about behavior change opportunities for bring about. For instance, the self-determination self-improvement [12], [13]. The desirable theory [17] suggests that activities which provide behaviors are facilitated through goal setting, an intrinsically motivating experience, or that are reminders, and goal achievement [13], [14]. Self- well-internalized (due to the perceived value of tracking experiences facilitate informational the activity or congruence with one’s values) can feedback and simultaneously also bring about lead to enhanced subjective wellbeing. An hedonic and affective responses [2], [15]. The indicator of intrinsic motivation that is widely informational feedback provided by self-tracking used in literature [27], [28] is perceived experiences help users realize the utilitarian value enjoyment and interest. Enjoyment and interest and perceived usefulness of the activity, which in reflect the users’ emotional response to the turn acts as a motivational tool [2], [16]. When intervention. On the other hand, perceived individuals recognize and identify the perceived usefulness reflects the users’ cognitive response value of an activity, they internalize and integrate to the intervention based on the utilitarian value of the desired behavior, yielding self-motivation and the experience. Perceived usefulness facilitates enhanced subjective wellbeing [17]. internalization and integration of extrinsically Gamification defined as ‘the use of game motivated behaviors [17], [29]. The advantages of design elements in non-game contexts’ [18, p. 1] internalization include more autonomous and is commonly integrated with self-tracking volitional commitment towards the desired technologies to enhance the intervention’s behavior and enhanced subjective wellbeing [17]. intended effects and promote engagement [4], [5]. Based on the theoretical underpinnings and Gamification serves a dual-purpose, users can literature presented in this section, we postulate derive both hedonic and utilitarian benefits [19], that the use of self-tracking technologies and [20]. The intrinsically motivating positive gamification can enhance wellbeing by eliciting experience that gamification is intended to positive emotional and cognitive responses based provide supports the initiation, reinforcement, and on hedonic and utilitarian benefits respectively. maintenance of healthy behaviors [4]. The Furthermore, we posit that gamified (relative to a hedonic design of gamified systems offers the non-gamified) self-tracking experience results in potential to generate a positive affective stronger emotional and cognitive responses, and experience that enhances the users’ perceived as a result enhanced wellbeing. Specifically, the benefits and sustain continued usage of self- data gathered through this study tests the tracking technologies [4]. The use of gamification following hypotheses: is known to evoke affective experiences [2], [3] H1: The use of gamification enhances the and satisfy intrinsic needs [21]. The positive effect on wellbeing (relative to a non-gamified experience provided through the use of self-tracking experience) gamification could potentially have a direct H2: The use of gamification results in stronger contribution to wellbeing [4]. emotional and cognitive responses (relative to a Wellbeing is defined as ‘a person’s cognitive non-gamified self-tracking experience) and affective evaluations of his or her life’ [22, p. H3: Enjoyment and interest enhance wellbeing 187]. Extant literature provides evidence that gain leisure-time physical activity is correlated with H4: Perceived usefulness of the experience positive affect, life satisfaction [23], and enhances wellbeing gain happiness [24], [25]. The magnitude of this association is small [23]. The literature has 2.2. Empirical evidence identified that subjective wellbeing is also affected by other factors, including the individual’s health (physical and mental), the Justifiably, the effectiveness of these individual’s lifecycle stage, income and motivational technologies needs to be employment, education, relationship status, corroborated with a body of empirical evidence supporting the promising beneficial effects [7], 120 [30], [31]. Notwithstanding the popularity of implementation of a four-week behavioral fitness trackers and gamification in industry intervention of physical activity, namely self- practice, empirical evidence supporting the tracking of physical activity, alone and in positive claims on wellbeing is scarce [4], [6]–[8]. conjunction with gamification. Extant empirical evidence on the effect of self- tracking of physical activity on wellbeing 3.2. Participants provides mixed evidence [32]. The use of self- tracking technology was found to be effective in The study was conducted amongst academic improving the individuals’ quality of life and wellbeing in corporate wellness programs [33], researchers and post-graduate research students at amongst older adults [34] and breast cancer the University of Malta. Participants were survivors [35]. Other literature [8] reports that recruited using a non-probabilistic convenience self-tracking experiences of physical activity had sampling method. Following an email invitation a statistically significant small positive effect on and a post on social media, interested participants the users’ perceived physical health and the sense were invited to review the information about the of goal accomplishment. Albeit positive, the study (including its objectives, duration, and requirements) and provide informed consent. increase reported for overall psychological Participants were eligible for this study if they wellbeing was not significant [8]. Likewise, were over 18 years of age, did not use a fitness another randomized controlled study [36] also reported that exercise-related self-tracking and tracker or a wearable to monitor their physical step goals did not substantially influence activity during the previous year, and had no psychological wellbeing. There is also evidence health issues (such as heart condition, chest pain, which suggests that while self-tracking can bone or joint pain, or dizziness) that they are increase the task performance, it may aware of, which could prevent them from simultaneously have negative effects on engaging in physical activity. Participants were ineligible if they were currently pregnant or have subjective wellbeing, including happiness and satisfaction by undermining the intrinsic been told by their doctor not to engage in physical motivation and enjoyment of performing such exercise. activities [37]. Thus, the effect of self-tracking of physical activity on the users’ wellbeing needs to 3.3. Data collection be further investigated. The majority of existing studies in the field of Data collection was carried out through self- gamification of physical activity, health and completed pen-and-paper questionnaires. In total, wellbeing [4], [30], [38] report positive effects on eighty participants completed both the pre- and user experience, affect, cognition and behavior, post-intervention surveys. which can have a positive impact on wellbeing Subjective wellbeing was measured using two [4]. However, despite the possibility of having a validated items identified in literature [42], positive impact on wellbeing, there is scant namely life satisfaction and happiness. Both items literature [39], [40] investigating whether were measured using an eleven-point Likert scale gamification of physical activity enhances the validated in previous empirical work [42]. The individual’s quality of life and wellbeing. A emotional response was measured in terms of the gamified community-wide physical activity users’ enjoyment and interest, while the cognitive intervention [41] reported increases in both self- response was measured in terms of the perceived reported physical activity and mental wellbeing. usefulness of the activity. The latter two Findings from other empirical studies [39], [40] constructs were measured using validated sub- reveal that whilst gamification led to an increase scales of the Intrinsic Motivation Inventory [27] in physical activity, there was no change on the on a seven-point Likert scale. Self-reported data quality of life or wellbeing measures reported. on potential predictors of wellbeing identified in literature [26] was also gathered. This includes 3. Materials and method demographic and lifestyle data including the self- reported stage of physical activity based on the 3.1. Study design transtheoretical model [43]. This study involved a two-wave longitudinal survey conducted before and after the 121 3.4. Procedure and interventions using a gamified platform (pointagram.com) that was accessible to all participants through an application installed on their smartphone or Following the eligibility screening criteria, a through a web browser. Visual images of the unique reference number was assigned to all gamified experiences are presented in the participants to ensure anonymity all throughout Supplementary Material. the study. Using an online random sequence generator (random.org), eligible participants (n = 80) who provided informed consent were 3.5. Statistical data analysis randomly assigned to either a non-gamified self- tracking experience (n = 20) or a gamified self- Statistical analysis was carried out in three tracking experience (n = 60). Participants were phases. First, descriptive statistics were computed blinded to group allocation and groups were for all the variables measured. Wilcoxon signed- color-coded to hide the identity of each group rank tests were computed to determine whether from participants. happiness and life satisfaction scores increased During the intervention period, all participants post-intervention compared to pre-intervention were given a smartwatch (Xiaomi Mi Band) to levels. The effect size r was computed using the Z track their physical activity. Earlier studies [44], value resulting from Wilcoxon test and the [45] show that these wearable devices are number of observations in the sample [50]. The adequately reliable in tracking physical activity, change for each wellbeing measure (life and hence these were preferred against other satisfaction and happiness) was computed as brands of pedometers due to their cost and battery follows (Equation 1): lifespan. All participants were instructed on how 𝐺𝑎𝑖𝑛 = 𝑝𝑜𝑠𝑡 − 𝑝𝑟𝑒 (1) to pair and sync the smartwatch with the where pre and post refers to the life corresponding mobile application, and to wear the satisfaction and happiness measures assessed device at all times. before and after the intervention. During the set-up of the wearables and the To test hypothesis H1, Mann-Whitney U tests corresponding application installed on their were carried out to determine whether the use of smartphones, all participants were allowed to gamification led to significantly higher gains in choose a personalized daily step target. Goal wellbeing measures. To increase the robustness of setting is a commonly used feature in self-tracking the results, an ANCOVA was carried out to motivational technologies [46] that supports determine whether there is a statistically users’ intrinsic motivation and self-regulation significant difference in the post-intervention [47]. Participants assigned to the non-gamified wellbeing scores between the non-gamified self- self-tracking group monitored whether they tracking group and the gamified self-tracking achieved their personal daily step goal target set group, after controlling for the pre-intervention on their smartwatch. wellbeing scores. Furthermore, we also tested for In addition to self-tracking, participants significant differences between the different assigned to a gamified experience were randomly gamification experiences and non-gamified self- assigned to either a group cooperation challenge tracking group in terms of wellbeing gains using (cooperative design), an individual competition Kruskal-Wallis test. (competitive design), or an inter-team Second, the constructs’ reliability for competition (competitive-cooperative design).2 Enjoyment and Interest and Perceived Usefulness The design of these gamified experiences was were measured using Cronbach's alpha (α), guided by the classification of gamification composite reliability (CR), and average variance features [48] and gamification design frameworks extracted (AVE). All convergent validity metrics [19], [49] identified in literature. The game obtained were checked against the thresholds elements utilized within these interventions are (Cronbach's α > 0.7, CR > 0.7 and AVE > 0.5) associated with the motivational constructs of the suggested in literature [51]. In order to test self-determination theory [17] to afford an hypothesis H2, we conducted Mann-Whitney U appealing and motivating experience for the users tests to determine whether the use of gamification [49]. The gamification experiences were designed led to significantly higher emotional and 2 These interventions were part of an experimental study examining the effect of different types of gamification designs on psychological and behavioral outcomes, detailed in a forthcoming publication [54]. 122 cognitive responses (relative to a non-gamified 4. Results self-tracking experience). Furthermore, we also tested for significant differences between the 4.1. Sample characteristics different gamification experiences and non- gamified self-tracking group in terms of the Descriptive statistics were computed for the emotional and cognitive responses using Kruskal- sample characteristics, including demographic Wallis tests. and lifestyle data, and the self-reported stage of Third, pair-wise bivariate correlations were physical activity at baseline (pre-intervention). computed to examine whether there is a The sample characteristics are set out in Table 1.3 relationship between wellbeing gains and potential predictors of change including Table 1 enjoyment and interest, perceived usefulness, Sample characteristics gamification and the baseline levels of life Variables % satisfaction and happiness. While correlation Male 44% analysis provides an insight on the strength of Female 56% positive or negative associations between these Young adult (20 - 34 years) 52% wellbeing constructs, and between them and their Middle aged (35 - 54 years) 45% potential predictors of change, it is not possible to Older adult (55+ years) 3% parse out the net effect of the latter variables on Full-time employed 65% the dependent measures. Thus, to test Hypotheses Have children under 16 years 22% H3 and H4, we carried out a multi-variate Have a steady relationship 72% regression analysis for each wellbeing measure Have sufficient income 88% (life satisfaction and happiness) to examine the contribution of each potential predictor of change Do voluntary work 23% and identify which factors were causing an effect Participate in religious/spiritual activity 31% on subjective wellbeing. The model used for this Participate in artistic/creative activity 20% analysis is presented below (Equation 2): Spend time in nature 61% Spend time with friends and family 96% WB_Gain = β0 + β1BaselineWB+ β2 Maintain a balance between work and 50% Enjoyment_Interest + β3Perceived_Usefulness + play + β4Gamification (2) Regular physical exercise 25% where WB_Gain is the dependent variable relating to the gain reported in life satisfaction and 4.2. Change in wellbeing happiness. The independent variables included are enjoyment and interest, perceived usefulness, Table 2 a dummy variable for gamification and the Pre- and post-intervention wellbeing scores baseline scores for life satisfaction and happiness. Pre Post Gain In order to increase the robustness of our findings, Mean Mean Mean Variables we computed multi-variate regression models on (SD) (SD) (SD) the post-intervention wellbeing measures (as 6.90 7.21 0.31 dependent variables), controlling for the Happiness (1.769) (1.998) (1.688) demographic and lifestyle variables, in addition to Life 6.86 7.34 0.48 the independent variables listed in Equation 2. Satisfaction (1.941) (1.916) (1.467) Statistical analysis was carried out using STATATM (version 16.1, StataCorp). Regression The findings show that there is a significant models were estimated using a robust estimator of increase in happiness (z = -2.298, p = 0.022, variance. effect size r = -0.182) and life satisfaction (z = - 2.911, p = 0.004, effect size r = -0.230) when 3 The distribution of participants between the gamified and the non- included as a covariate amongst other variables in the multi-variate gamified self-tracking groups was relatively well-balanced in terms regression model analyzing the potential predictors of wellbeing of all demographic and lifestyle characteristics, with the exception of change. The results presented as part of the Supplementary Material having less participants in the non-gamified group who had children confirm that having children under the age of sixteen was not a under the age of sixteen, even though randomization was employed. significant predictor to the change reported in wellbeing measures. For sensitivity analysis, ‘having children under 16 years of age’ was 123 comparing post-intervention scores to pre- between the groups in terms of reported intervention scores (see Table 2). enjoyment and interest (U = 513.50, z = -0.988, p Non-gamified and gamified self-tracking = 0.323) and perceived usefulness (U = 509.00, z groups reported similar increases in wellbeing = -1.017, p = 0.309), thus rejecting Hypothesis 2 measures (see Table 3). The findings show that (H2: The use of gamification results in stronger the use of gamification did not produce emotional and cognitive responses relative to a significantly higher gains in happiness (U = non-gamified self-tracking experience). 587.5; z = -0.143, p = 0.886) and life satisfaction The analysis also confirms that enjoyment and (U = 529.0; z = -0.816, p = 0.414) relative to a interest (χ2(3) = 1.160, p = 0.657), and perceived non-gamified self-tracking experience, thus usefulness (χ2(3) = 1.969, p = 0.579) were not rejecting Hypothesis 1 (H1: The use of statistically significantly different between the gamification enhances the effect on wellbeing different gamification experiences and the non- relative to a non-gamified self-tracking gamified self-tracking group. experience). Kruskal-Wallis tests also confirm that there are no significant differences in the Table 4 gains reported for happiness (χ2(3) = 1.944, p = The emotional and cognitive responses for the 0.584) and life satisfaction (χ2(3) = 3.066, p = non-gamified and gamified self-tracking groups 0.381) between the different gamified experiences Non-Gamified Gamified and non-gamified self-tracking experience. Variables Mean (SD) Mean (SD) ANCOVA results show that after adjusting for the Enjoyment & pre-test wellbeing scores, there are no statistically 6.40 (0.746) 6.16 (0.901) Interest significant differences in the post-intervention wellbeing scores of the non-gamified self- Perceived 5.28 (1.186) 5.52 (1.324) tracking group and the gamified group for either Usefulness happiness (F(1,77) = 0.029, p = 0.865) or life satisfaction (F(1,77) = 0.140, p = 0.709). 4.4. Predictors of wellbeing change Table 3 The intercorrelations between the gains Wellbeing gain for the non-gamified and gamified reported in wellbeing outcomes and the variables self-tracking groups hypothesized to cause an increase in subjective Non-Gamified Gamified wellbeing are presented in the Supplementary Variables Mean (SD) Mean (SD) Material. The correlations indicate a significant Happiness Gain 0.35 (1.899) 0.30 (1.629) positive relationship between happiness gain and Life Satisfaction life satisfaction gain, a significant negative 0.35 (1.872) 0.52 (1.321) correlation with baseline happiness and life Gain satisfaction scores, and a significant positive association with the users’ enjoyment and interest, 4.3. Emotional and cognitive and perceived usefulness. responses The results of the multi-variate regression analysis (see Table 5) provide evidence to the The subscales used to measure the users’ predictors of the gains reported in happiness and emotional and cognitive responses were found to life satisfaction. The emotional psychological be reliable, indicating internal consistency among response to the intervention measured through the the scale items used to measure each specific individuals’ enjoyment and interest (hedonic construct. Enjoyment and Interest sub-scale (α = benefit) produced a significant positive effect (β = 0.735; CR = 0.859; AVE = 0.677) resulted in a 0.596) that increased the individuals’ happiness scale with M = 6.22 and SD = 0.867, and the levels, supporting Hypothesis 3 (H3: Enjoyment Perceived Usefulness sub-scale (α = 0.808; CR = and interest enhance wellbeing gain). The 0.891; AVE = 0.734) resulted in a scale with M = cognitive psychological response to the 5.46 and SD = 1.288. intervention measured through the perceived Both the non-gamified and gamified self- usefulness (utilitarian benefit) produced a tracking groups reported similar positive significant positive effect (β = 0.450) that psychological outcomes (see Table 4). Results increased the individuals’ life satisfaction levels, show that there are no significant differences supporting Hypothesis 4 (H4: Perceived 124 usefulness of the experience enhances wellbeing measured as enjoyment and interest (hedonic gain). The use of gamification did not produce a benefit) and as perceived usefulness (utilitarian significant positive effect, thus providing further benefit). Previous research [2], [20] highlights the evidence to Hypothesis H1. The results also importance of both perceived usefulness and provide evidence that the baseline measure of enjoyment for continued use of motivational happiness was a significant predictor to the information systems. happiness gain (β = -0.331). The negative Previous empirical studies [39], [40] indicated coefficient value for baseline happiness indicates that physical activity interventions involving that lower happiness levels at baseline contributed activity trackers and gamification did not produce to higher happiness gains. Similarly, the baseline a significant change in wellbeing and quality of measure of life satisfaction was a significant life measures. In contrast, our findings suggest predictor to the gain reported in life satisfaction (β that experiences of self-tracking and gamification = -0.328), meaning that lower life satisfaction have statistically significant positive effects on levels at baseline contributed to higher life happiness (effect size r = -0.182) and life satisfaction gains. To increase the robustness of satisfaction (effect size r = -0.230). These effects our findings, we controlled for the demographic corroborate the standardized effect sizes observed and lifestyle variables. We find that the results in previous literature following the use of self- remain unchanged (see Supplementary Material). tracking technologies [8], [35]. Gamification is commonly integrated with Table 5 self-tracking technologies to promote engagement Multi-variate regression on wellbeing gains and enhance the intervention’s intended effects Variables Happiness Life [4], [5]. Nevertheless, the findings from our study Gain Satisfaction show that at the end of the intervention (after four Gain weeks), gamified and non-gamified self-tracking Baseline -0.331*** evoked similar positive emotional and cognitive responses, yielding similar wellbeing gains. Happiness (0.113) Literature suggests that enjoyment and perceived Baseline Life -0.328*** usefulness of gamification declines with use [52]. Satisfaction (0.086) Thus, future work could consider more frequent Enjoyment & 0.596** 0.295 measurements during the intervention period. Interest (0.281) (0.222) The gains in wellbeing measures were Perceived 0.176 0.450*** attributed to the users’ positive psychological Usefulness (0.174) (0.132) responses resulting from gamified and non- Gamification 0.029 0.094 gamified self-tracking experiences of physical (0.413) (0.364) activity. Specifically, enjoyment and interest were Constant -2.093 -1.631 linked to an increase in the individuals’ happiness (1.619) (1.454) levels. In turn, the perceived usefulness of the experience was associated with an increase in the Observations 80 80 individuals’ life satisfaction levels. Our findings R-squared 0.279 0.437 support existing literature suggesting that intrinsic F value F(4, 75) = 6.29 F(4, 75) = 9.71 motivation and autonomous forms of extrinsic P value p < 0.001 p < 0.001 motivation enhance wellbeing [17]. In synthesis, Robust standard errors in parentheses the findings suggest that the hedonic benefit of the *** p<0.01, ** p<0.05, * p<0.1 experience enhances happiness levels (hedonic wellbeing), while the utilitarian benefit of the experience enhances life satisfaction levels 5. Discussion and conclusion (eudaimonic wellbeing) [53]. Our findings provide insights into how The findings support the theoretical prediction subjective wellbeing is influenced by self- that experiences of self-tracking, alone and in tracking technologies and the use of gamification, conjunction with gamification elicit positive an area which is underexplored in literature. Yet, emotional and cognitive responses that yield despite our contributions, we acknowledge that enhanced wellbeing. Both gamified and non- there are some limitations which could be gamified self-tracking experiences facilitated addressed in future research. First, this study was similar positive psychological responses, conducted amongst academic members and post- 125 graduate students. Future studies should also gamification, quantified-self and social examine a broader population and a larger sample networking,” Int J Inf Manage, vol. 46, pp. size with longer timeframes to increase the 151–162, Jun. 2019, doi: generalizability and robustness of the findings. 10.1016/j.ijinfomgt.2018.12.004. Understandably, there are challenges to conduct [3] J. Hamari, L. Hassan, and A. Dias, empirical studies involving wearable physical “Gamification, quantified-self or social activity trackers with a large sample size. networking? Matching users’ goals with However, the accumulation of knowledge from motivational technology,” User Model rigorous empirical work on the effect of self- User-adapt Interact, vol. 28, no. 1, pp. 35– tracking and gamification on health-related 74, Mar. 2018, doi: 10.1007/s11257-018- behaviors and societal wellbeing would have 9200-2. practical relevance. Second, this study compared [4] D. Johnson, S. Deterding, K.-A. Kuhn, A. longitudinal wellbeing data of gamified and non- Staneva, S. Stoyanov, and L. Hides, gamified self-tracking experiences. Future studies “Gamification for health and wellbeing: A could also include an inactive control group, with systematic review of the literature,” no access to physical activity trackers. Third, this Internet Interv, vol. 6, pp. 89–106, Nov. study examined wellbeing measures pre- and 2016, doi: 10.1016/j.invent.2016.10.002. post-intervention. Future studies could consider [5] V. W. S. Cheng, T. Davenport, D. gathering data on the users’ experience and Johnson, K. Vella, and I. B. Hickie, wellbeing more frequently during the “Gamification in apps and technologies intervention, possibly using a diary research for improving mental health and well- approach or through real-time customer being: Systematic review,” J Med Internet experience tracking. This would allow more Res, vol. 21, no. 6, pp. 1–15, 2019, doi: granular data on the users’ interaction with 10.2196/13717. motivational information systems and a more [6] S. Hermsen, J. Frost, R. J. Renes, and P. comprehensive view of the effect on the users’ Kerkhof, “Using feedback through digital experience and wellbeing. technology to disrupt and change habitual To conclude, results from this study behavior: A critical review of current demonstrate that using wearable fitness trackers literature,” Comput Human Behav, vol. 57, (with and without the use of gamification) pp. 61–74, Apr. 2016, doi: increases subjective wellbeing. The value co- 10.1016/j.chb.2015.12.023. created through such meaningful experiences [7] A. Rapp, F. Hopfgartner, J. Hamari, C. improves people’s quality of life and wellbeing. Linehan, and F. Cena, “Strengthening gamification studies: Current trends and 6. Declaration of interest future opportunities of gamification research,” International Journal of Human Computer Studies, vol. 127, no. November The authors declare that they have no conflict 2018, pp. 1–6, 2019, doi: of interest. This research did not receive any 10.1016/j.ijhcs.2018.11.007. specific grant from funding agencies in the public, [8] B. Stiglbauer, S. Weber, and B. Batinic, commercial, or not-for-profit sectors. “Does your health really benefit from using a self-tracking device? Evidence 7. References from a longitudinal randomized control trial,” Comput Human Behav, vol. 94, no. [1] A. Neupane, D. Hansen, A. Sharma, J. A. January, pp. 131–139, 2019, doi: Fails, B. Neupane, and J. Beutler, “A 10.1016/j.chb.2019.01.018. Review of Gamified Fitness Tracker Apps [9] D. Lupton, The Quantified Self. John and Future Directions,” in Proceedings of Wiley & Sons., 2016. the Annual Symposium on Computer- [10] I. Li, A. Dey, and J. Forlizzi, “A stage- Human Interaction in Play, New York, based model of personal informatics NY, USA: ACM, Nov. 2020, pp. 522–533. systems,” in Proceedings of the 28th doi: 10.1145/3410404.3414258. international conference on Human [2] L. Hassan, A. Dias, and J. Hamari, “How factors in computing systems - CHI ’10, motivational feedback increases user’s New York, New York, USA: ACM Press, benefits and continued use: A study on 126 2010, p. 557. doi: Conference Extended Abstracts on Human 10.1145/1753326.1753409. Factors in Computing Systems (CHI [11] A. Rapp, A. Marcengo, L. Buriano, G. 2011), pp. 2425–2428, 2011, doi: Ruffo, M. Lai, and F. Cena, “Designing a 10.1145/1979742.1979575. personal informatics system for users [19] D. Liu, R. Santhanam, and J. Webster, without experience in self-tracking: a case “Toward Meaningful Engagement: A study,” Behaviour & Information Framework for Design & Research of Technology, vol. 37, no. 4, pp. 335–366, Gamified Information Systems,” MIS Apr. 2018, doi: Quarterly, vol. 41, no. 4, pp. 1011–1034, 10.1080/0144929X.2018.1436592. 2017. [12] A. Bandura, “Social cognitive theory of [20] D. B. Köse, B. Morschheuser, and J. self-regulation,” Organ Behav Hum Decis Hamari, “Is it a tool or a toy? How user Process, vol. 50, no. 2, pp. 248–287, Dec. conceptions of a system’s purpose affect 1991, doi: 10.1016/0749-5978(91)90022- their experience and use,” Int J Inf L. Manage, vol. 49, no. June, pp. 461–474, [13] E. T. Kersten-van Dijk, J. H. D. M. 2019, doi: Westerink, F. Beute, and W. A. 10.1016/j.ijinfomgt.2019.07.016. IJsselsteijn, “Personal Informatics, Self- [21] N. Xi and J. Hamari, “Does gamification Insight, and Behavior Change: A Critical satisfy needs? A study on the relationship Review of Current Literature,” Hum between gamification features and Comput Interact, vol. 32, no. 5–6, pp. 268– intrinsic need satisfaction,” Int J Inf 296, Nov. 2017, doi: Manage, vol. 46, pp. 210–221, Jun. 2019, 10.1080/07370024.2016.1276456. doi: 10.1016/j.ijinfomgt.2018.12.002. [14] S. a Munson and S. Consolvo, “Exploring [22] E. Diener, S. Oishi, and R. E. Lucas, goal-setting, rewards, self-monitoring, and “Subjective Well-Being: The Science of sharing to m motivate physical activity,” Happiness and Life Satisfaction,” in The 6th International Conference on Pervasive Oxford Handbook of Positive Psychology, Computing Technologies for Healthcare S. J. Lopez and C. R. Snyder, Eds., Oxford (PervasiveHealth), pp. 25–32, 2012. University Press, 2009, pp. 186–194. doi: [15] L. Hassan, N. Xi, B. Gurkan, J. Koivisto, 10.1093/oxfordhb/9780195187243.013.0 and J. Hamari, “Gameful Self-Regulation: 017. A Study on How Gamified Self-Tracking [23] C. W. Wiese, L. Kuykendall, and L. Tay, Features Evoke Gameful Experiences,” “Get active? A meta-analysis of leisure- Proceedings of the 53rd Hawaii time physical activity and subjective well- International Conference on System being,” J Posit Psychol, vol. 13, no. 1, pp. Sciences, no. January, 2020, doi: 57–66, Jan. 2018, doi: 10.24251/hicss.2020.138. 10.1080/17439760.2017.1374436. [16] R. M. Ryan and E. L. Deci, “Intrinsic and [24] M. Argyle, The Psychology of Happiness, extrinsic motivation from a self- 2nd Editio. London: Routledge, 2001. doi: determination theory perspective: 10.4324/9781315812212. Definitions, theory, practices, and future [25] Z. Zhang and W. Chen, “A Systematic directions,” Contemp Educ Psychol, vol. Review of the Relationship Between 61, no. April, 2020, doi: Physical Activity and Happiness,” J 10.1016/j.cedpsych.2020.101860. Happiness Stud, vol. 20, no. 4, pp. 1305– [17] R. M. Ryan and E. L. Deci, “Self- 1322, 2019, doi: 10.1007/s10902-018- determination theory and the facilitation of 9976-0. intrinsic motivation, social development, [26] M. Briguglio, “Perspectives on and well-being.,” American Psychologist, Wellbeing,” Brill | Sense, Mar. 2019. doi: vol. 55, no. 1, pp. 68–78, 2000, doi: 10.1163/9789004394179. 10.1037/0003-066X.55.1.68. [27] R. M. Ryan, “Control and information in [18] S. Deterding, D. Dixon, R. Khaled, L. the intrapersonal sphere: An extension of Nacke, M. Sicart, and K. O’Hara, cognitive evaluation theory.,” J Pers Soc “Gamification: Using Game Design Psychol, vol. 43, no. 3, pp. 450–461, 1982, Elements in Non-Game Contexts,” doi: 10.1037/0022-3514.43.3.450. Proceedings of the 2011 Annual 127 [28] J. Wu and X. Lu, “Effects of extrinsic and [36] L. Busch, T. Utesch, P.-C. Bürkner, and B. intrinsic motivators on using utilitarian, Strauss, “The Influence of Fitness-App hedonic, and dual-purposed information Usage on Psychological Well-Being and systems: A meta-analysis,” J Assoc Inf Body Awareness—A Daily Diary Syst, vol. 14, no. 3, pp. 153–191, 2013, Randomized Trial,” J Sport Exerc doi: 10.17705/1jais.00325. Psychol, vol. 42, no. 3, pp. 249–260, Jun. [29] E. L. Deci, H. Eghrari, B. C. Patrick, and 2020, doi: 10.1123/jsep.2019-0315. D. R. Leone, “Facilitating Internalization: [37] J. Etkin, “The hidden cost of personal The Self‐Determination Theory quantification,” Journal of Consumer Perspective,” Journal of Personality, vol. Research, vol. 42, no. 6, pp. 967–984, 62, no. 1. pp. 119–142, 1994. doi: 2016, doi: 10.1093/jcr/ucv095. 10.1111/j.1467-6494.1994.tb00797.x. [38] M. Schmidt-Kraepelin, S. Warsinsky, S. [30] J. Koivisto and J. Hamari, “The rise of Thiebes, and A. Sunyaev, “The role of motivational information systems: A gamification in health behavior change: A review of gamification research,” Int J Inf Review of Theory-driven Studies,” Manage, vol. 45, no. October 2018, pp. Proceedings of the Annual Hawaii 191–210, 2019, doi: International Conference on System 10.1016/j.ijinfomgt.2018.10.013. Sciences, vol. 2020-Janua, no. January, pp. [31] J. Hamari, J. Koivisto, and H. Sarsa, “Does 1256–1265, 2020, doi: gamification work? - A literature review of 10.24251/hicss.2020.155. empirical studies on gamification,” in [39] C. Maher et al., “A web-based, social Proceedings of the Annual Hawaii networking physical activity intervention International Conference on System for insufficiently active adults delivered Sciences, IEEE Computer Society, 2014, via Facebook app: Randomized controlled pp. 3025–3034. trial,” J Med Internet Res, vol. 17, no. 7, [32] D. Jin, H. Halvari, N. Maehle, and A. H. 2015, doi: 10.2196/jmir.4086. Olafsen, “Self-tracking behaviour in [40] L. Paul et al., “Increasing physical activity physical activity: a systematic review of in stroke survivors using STARFISH, an drivers and outcomes of fitness tracking,” interactive mobile phone application: a Behaviour and Information Technology, pilot study,” Top Stroke Rehabil, vol. 23, vol. 41, no. 2, pp. 242–261, 2022, doi: no. 3, pp. 170–177, Apr. 2016, doi: 10.1080/0144929X.2020.1801840. 10.1080/10749357.2015.1122266. [33] L. Giddens, D. Leidner, and E. Gonzalez, [41] M. A. Harris, “The relationship between “The role of fitbits in corporate wellness physical inactivity and mental wellbeing: programs: Does step count matter?,” Findings from a gamification-based Proceedings of the Annual Hawaii community-wide physical activity International Conference on System intervention,” Health Psychol Open, vol. Sciences, vol. 2017-Janua, pp. 3627–3635, 5, no. 1, 2018, doi: 2017, doi: 10.24251/hicss.2017.438. 10.1177/2055102917753853. [34] A. Suh and M. Li, “How the use of mobile [42] European Social Survey, “Measuring and fitness technology influences older adults’ Reporting on Europeans’ Wellbeing: physical and psychological well-being,” Findings from the European Social Comput Human Behav, vol. 131, no. Survey,” 2018. September 2021, p. 107205, 2022, doi: https://www.europeansocialsurvey.org/da 10.1016/j.chb.2022.107205. ta/download.html?r=9 [35] J. K. H. Vallance, K. S. Courneya, R. C. [43] J. O. Prochaska and W. F. Velicer, “The Plotnikoff, Y. Yasui, and J. R. Mackey, Transtheoretical Model of Health “Randomized Controlled Trial of the Behavior Change,” American Journal of Effects of Print Materials and Step Health Promotion, vol. 12, no. 1, pp. 38– Pedometers on Physical Activity and 48, Sep. 1997, doi: 10.4278/0890-1171- Quality of Life in Breast Cancer 12.1.38. Survivors,” Journal of Clinical Oncology, [44] M. K. Tam and S. Y. Cheung, “Validation vol. 25, no. 17, pp. 2352–2359, Jun. 2007, of consumer wearable activity tracker as doi: 10.1200/JCO.2006.07.9988. step measurement in free-living conditions,” Finnish Journal of eHealth 128 and eWelfare, vol. 11, no. 1–2, pp. 68–75, no. 1, pp. 141–166, Feb. 2001, doi: 2019, doi: 10.23996/fjhw.76673. 10.1146/annurev.psych.52.1.141. [45] J. Xie, D. Wen, L. Liang, Y. Jia, L. Gao, [54] E. M. Grech, M. Briguglio, and E. Said, and J. Lei, “Evaluating the validity of “Can Gamification Help with Physical current mainstream wearable devices in Inactivity? A Randomised Controlled fitness tracking under various physical Field Experiment on the Psychological activities: Comparative study,” JMIR and Behavioural Effects of Gamification,” Mhealth Uhealth, vol. 6, no. 4, 2018, doi: [Manuscript submitted for publication], 10.2196/mhealth.9754. 2023. [46] N. Aldenaini, F. Alqahtani, R. Orji, and S. Sampalli, “Trends in Persuasive Technologies for Physical Activity and Sedentary Behavior: A Systematic Review,” Front Artif Intell, vol. 3, no. April, 2020, doi: 10.3389/frai.2020.00007. [47] G. P. Latham and E. A. Locke, “Self- regulation through goal setting,” Organ Behav Hum Decis Process, vol. 50, no. 2, pp. 212–247, Dec. 1991, doi: 10.1016/0749-5978(91)90021-K. [48] B. Morschheuser, A. Maedche, and D. Walter, “Designing Cooperative Gamification: Conceptualization and Prototypical Implementation,” in Proceedings of the 2017 ACM Conference on Computer Supported Cooperative Work and Social Computing, New York, NY, USA: ACM, Feb. 2017, pp. 2410– 2421. doi: 10.1145/2998181.2998272. [49] J. Buckley, T. DeWille, C. Exton, G. Exton, and L. Murray, “A Gamification– Motivation Design Framework for Educational Software Developers,” Journal of Educational Technology Systems, vol. 47, no. 1, pp. 101–127, 2018, doi: 10.1177/0047239518783153. [50] R. Rosenthal, Meta-Analytic Procedures for Social Research. 2455 Teller Road, Thousand Oaks California 91320 United States of America: SAGE Publications, Inc., 1991. doi: 10.4135/9781412984997. [51] J. Hair, W. Black, B. Babin, and R. Anderson, Multivariate Data Analysis, Seventh Ed. New Jersey: Pearson Education, 2010. [52] J. Koivisto and J. Hamari, “Demographic differences in perceived benefits from gamification,” Comput Human Behav, vol. 35, pp. 179–188, 2014, doi: 10.1016/j.chb.2014.03.007. [53] R. M. Ryan and E. L. Deci, “On Happiness and Human Potentials: A Review of Research on Hedonic and Eudaimonic Well-Being,” Annu Rev Psychol, vol. 52, 129 Supplementary Material Description of the gamification experiences Group cooperation challenge Inter-team competition This group had a quest to reach a target step count This group had a team versus team competition. by the end of the week. The map shows a pirate Participants were randomised to teams of four making his way to reach the treasure chest with a players each. Accumulated points based on the countdown timer indicating the time left for the daily step counts were visible on a leaderboard participants to complete the challenge. The which indicated the ranking of all the teams. progress that the pirate made towards the treasure Virtual trophies were awarded to the top three chest reflected the users’ accumulated points teams with the highest step counts. based on their step counts. Figure S1 Group cooperation challenge Individual competition Figure S3 This group had an individual competition Inter-team competition amongst the participants. Points based on their daily step counts were visible on a leaderboard which indicated the ranking of all the participants. Virtual trophies were awarded to the top three players with the highest step counts. Figure S2 Individual competition 130 Supplementary Tables Table S1 Correlations for the study variables Variables 1 2 3 4 5 6 7 1. Happiness gain -- 2. Life satisfaction gain 0.691** -- 3. Baseline happiness -0.333** -0.323** -- ** 4. Baseline life satisfaction -0.184 -0.395 0.881** -- ** ** 5. Enjoyment & Interest 0.393 0.428 0.015 0.041 -- 6. Perceived usefulness 0.321** 0.482** 0.076 0.083 0.696** -- 7. Gamification -0.013 0.050 -0.016 -0.026 -0.120 0.079 -- ** Correlation is significant at the 0.01 level (2-tailed). Table S2 Regression models for post-intervention happiness and life satisfaction including demographic and lifestyle variables Variables POST POST Happiness model Life Satisfaction model Baseline happiness 0.669*** (0.159) Baseline life satisfaction 0.743*** (0.096) Enjoyment & Interest 0.704** (0.337) 0.335 (0.214) Perceived usefulness 0.101 (0.233) 0.508*** (0.162) Gamification 0.079 (0.420) 0.123 (0.372) Male gender 0.160 (0.349) 0.132 (0.316) Young adult -0.201 (0.329) -0.336 (0.298) Children under 16 years -0.358 (0.733) -0.353 (0.432) Voluntary work -0.148 (0.364) 0.125 (0.254) Religious participation 0.208 (0.474) -0.376 (0.281) Artistic activity -0.322 (0.628) -0.119 (0.364) Spends time in nature -0.120 (0.381) -0.418 (0.312) Spends time with family & friends 1.208 (1.140) 0.280 (0.567) Balance work and play -0.380 (0.332) -0.490** (0.239) Regular physical exercise 0.074 (0.467) 0.367 (0.351) Constant -3.156 (1.981) -2.263* (1.332) Observations 80 80 R-squared 0.521 0.718 Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 131