This is an open-access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in JMIR mhealth and uhealth, is properly cited. The complete bibliographic information, a link to the original publication on http://mhealth.jmir.org/, as well as this copyright and license information must be included.
Goal setting is among the most common behavioral change techniques employed in contemporary self-tracking apps. For these techniques to be effective, it is relevant to understand how the visual presentation of goal-related outcomes employed in the app design affects users’ responses to their self-tracking outcomes.
This study examined whether a spatially close (vs distant) presentation of mixed positive and negative self-tracking outcomes from multiple domains (ie, activity, diet) on a digital device’s screen can provide users the opportunity to hedonically edit their self-tracking outcome profile (ie, to view their mixed self-tracking outcomes in the most positive light). Further, this study examined how the opportunity to hedonically edit one’s self-tracking outcome profile relates to users’ future health behavior intentions.
To assess users’ responses to a spatially close (vs distant) presentation of a mixed-gain (vs mixed-loss) self-tracking outcome profile, a randomized 2×2 between-subjects online experiment with a final sample of 397 participants (mean age 27.4, SD 7.2 years; 71.5%, 284/397 female) was conducted in Germany. The experiment started with a cover story about a fictitious self-tracking app. Thereafter, participants saw one of four manipulated self-tracking outcome profiles. Variables of interest measured were health behavior intentions, compensatory health beliefs, health motivation, and recall of the outcome profile. We analyzed data using chi-square tests (SPSS version 23) and moderated mediation analyses with the PROCESS macro 2.16.1.
Spatial distance facilitated hedonic editing, which was indicated by systematic memory biases in users’ recall of positive and negative self-tracking outcomes. In the case of a mixed-gain outcome profile, a spatially close (vs distant) presentation tended to increase the underestimation of the negative outcome (
When spatial distance between the presentations of mixed self-tracking outcomes provided the opportunity to hedonically edit one’s self-tracking outcome profile, users recalled their self-tracking outcomes in a more positive light. Especially for users with lower health motivation, the opportunity to hedonically edit one’s mixed self-tracking outcome profile led to reduced health behavior intentions. To prevent the occurrence of hedonic editing in users’ responses to visually presented self-tracking outcome profiles, further research is necessary to determine the ideal distance that should be employed in the app design for the presentation of mixed self-tracking outcomes on a digital device’s screen.
Given the growing spread of wearable technological devices and mobile phones alongside with mobile apps, consumers nowadays increasingly practice digital forms of self-tracking [
So far, existing studies have examined the effectiveness of health-related mobile phone interventions and technology-enhanced interventions in general [
This lack of research is surprising, considering that users’ reflections of their self-tracking outcomes most likely will determine their health-related intentions and actions [
First, to apply principles of mental accounting—originally used to describe how consumers mentally code, categorize, and evaluate multiple economic outcomes [
Proposed model of how hedonic editing opportunity as a function of spatial distance and self-tracking outcome profile influences users’ health behavior intentions. The self-tracking outcome profile is a mixed gain if there is a positive outcome in one domain and a smaller negative outcome in the other domain. Conversely, in a mixed-loss outcome profile, there is a large negative outcome in one domain and a smaller positive outcome in the other domain.
Second, to describe how consumers mentally combine multiple outcomes within a single account, Thaler [
Third, to explain how consumers evaluate combinations of positive and negative outcomes, Thaler [
Regarding the influence of spatial distance on the use of hedonic editing strategies, we draw on previous research, which suggests that temporal distance between the occurrence of a positive and negative outcome facilitates hedonic editing [
Following the reasoning of Cowley [
Therefore, our first hypothesis is that, in the case of a mixed-gain outcome profile, a close (vs distant) presentation of self-tracking outcomes will lead to an underestimation (vs accurate estimation or overestimation) of the recalled small negative outcome and that, in the case of a mixed-loss outcome profile, a distant (vs close) presentation of self-tracking outcomes will lead to a more accurate (vs overestimation or underestimation) memory of the small positive outcome.
Because mental accounts are considered to function as self-regulatory mechanisms [
We suggest that this effect might be explained by a motivated cognitive justification mechanism. We argue that particularly users with a low as opposed to a high health motivation may—probably unconsciously—justify using the opportunity to hedonically edit their self-tracking outcomes, which allows them to have lower health behavior intentions in response to their tracking results. One way to cognitively justify reduced health behavior intentions in response to the hedonic editing opportunity is the formation of compensatory health beliefs, which are “beliefs that the negative effects of an unhealthy behavior can be compensated for, or ‘neutralized,’ by engaging in a healthy behavior” ([
This study aimed to establish that spatial distance in the presentation of self-tracking outcomes facilitates hedonic editing and examined whether a close compared to a distant presentation of self-tracking outcomes in the case of a mixed-gain or a mixed-loss outcome profile biased users’ memory of self-tracking outcomes as proposed by hedonic editing principles. Further, this study examined how the opportunity to hedonically edit one’s outcome profile affected users’ health behavior intentions. To investigate our hypotheses, we conducted a 2 (spatial distance: close vs distant) × 2 (outcome profile: mixed gain vs mixed loss) between-subjects online experiment.
The experiment was designed in the German language using SoSci Survey, a software package for conducting online surveys. We tested the technical functionality of the electronic questionnaire and the correctness of electronic data recording as well as the transmission of collected data before making the online experiment public. A convenience sample was used because the invitation link to the open online survey was distributed in various groups of social media networks and shared on two survey websites. Consequences of this procedure for the survey population with respect to the sex (ie, more females) and age (ie, relatively young) were expected. We recruited participants from December 2016 to January 2017. The study was approved by the Head of the Department of Retailing and Customer Management, University of Cologne, Germany, to confirm compliance with ethical standards. As such, before beginning of the survey, participants were informed on the entry page about the general topic, purpose, and procedure of the study, and that their participation was voluntary and the data would be treated anonymously. Participants had to give their consent by clicking on the button “I agree” to continue with the study. No incentives were offered.
The study employed a 2 (spatial distance: close vs distant) × 2 (outcome profile: mixed gain vs mixed loss) between-subjects design and questionnaires with the respective experimental condition were randomly displayed by the survey software. The cover story explained that the study’s aim was to optimize the design of a yet-unreleased self-tracking app that measures health-related information so that users can easily monitor their activity, diet, and overall energy balance. The instruction sheet stated that the study sought to find out how comprehensible the app layout was. We briefed participants about the integral elements of the fictitious self-tracking app and explained that they will see an illustration of self-tracking results as they were obtained at the end of a certain day, containing information about activity (calories burned from physical exercise), diet (calories consumed from food and drinks), and overall energy balance (difference between calorie burning and consumption). We told participants that the app sets daily caloric goals based on individual needs for the domain of activity (minimum amount of calories that should be burned from activity during the day) and diet (maximum amount of calories that should be consumed during the day). These goals should be achieved or surpassed to obtain an overall even energy balance, where calorie consumption does not exceed calorie burning. Thereafter, participants saw a fictive self-tracking outcome profile including the display of goal achievement in the domains of activity, diet, and overall energy balance. We asked participants to imagine that they themselves had obtained the depicted self-tracking outcomes at the end of a certain day.
We presented participants either a mixed-gain or mixed-loss self-tracking outcome profile with the outcomes in the domains of activity and diet being either spatially close to or distant from one another. The four different self-tracking outputs were designed with the graphic program Adobe InDesign CS6 (see
In the mixed-gain condition, the self-tracking results showed participants that they had underscored their caloric activity goal by 65 kcal and achieved their caloric diet goal by consuming 220 kcal less than the target amount of calories, hence having an overall positive energy balance. In the mixed-loss condition, the self-tracking results showed participants that they had exceeded their caloric activity goal by 65 kcal, but that they had failed to achieve their caloric dietary goal by consuming 540 kcal too much, hence having an overall negative energy balance. The exact goal value was not indicated to prevent participants from comparing the fictitious goal with their own needs or their own set goals. The outcomes of activity and nutrition were highlighted graphically through bar graphs and marked with plus or minus signs. The colors green and red were used for the plus and minus signs, respectively, and were consistently used in the output design to enable fast recognition of gains or losses. Spatial distance between self-tracking outcomes from the domain of activity and diet in the close condition was set to 94 pixels and in the distant condition to 667 pixels. Further, the output included information about the overall energy balance to prompt a mental connection between activity and nutrition.
Illustration of mixed-gain and mixed-loss self-tracking stimuli.
All measures took place immediately after participants had seen the self-tracking outcome profile. The questionnaire assessed participants’ health behavior intentions by asking how likely it was that they had activity-related and diet-related behavioral intentions in response to their self-tracking outcomes (four items on a 7-point scale anchored visually at 1=very unlikely and 7=very likely): “I want to be more active in the near future,” “I want to do more sports in the near future,” “I want to eat healthier in the near future,” and “I want to eat less high-caloric in the near future” (α=.70). Similar to previous work [
We analyzed our data using the statistical package of SPSS version 23 (IBM Corp, Armonk, NY, USA). All effects are reported as significant at
Second, to test our second hypothesis, we analyzed whether users’ levels of health motivation moderated the direct effect of self-tracking outcome presentation that either does or does not provide the opportunity for hedonic editing on health behavior intentions as well as the indirect effect of self-tracking outcome presentation on health behavior intentions through compensatory health beliefs. To simultaneously test both the moderator and mediator, we used the PROCESS macro version 2.16.1 for SPSS, which is suited for conducting a moderated mediation analysis (moderated mediation; model 8 [
A total of 584 participants completed the online experiment. Manipulation checks served to exclude 187 participants who indicated they remembered a wrong outcome profile as well as participants who indicated they remembered an extremely high or low deviation from the determined caloric goals in the domain of activity (kcal values ≤10 and ≥450) or diet (kcal values ≤10 and ≥1000) in the free recall question on their goal deviation (kcal value). The final sample consisted of 397 participants with cell sizes ranging between n=84 and n=113. The sample consisted of 71.5% (284/397) female participants and the mean age was 27.4 (SD 7.2) years. In terms of the highest educational level attained, of the 397 participants, 2 (0.5%) had a lower secondary school leaving certificate, 33 (8.3%) had an intermediate or general secondary school leaving certificate, 154 (38.8%) had a general or subject-linked higher education entrance qualification, 133 (33.5%) had a bachelor’s degree, 73 (18.4%) had a master’s degree or a diploma, and 2 (0.5%) had a doctoral degree. We used participants’ voluntary information about body weight (in kg) as well as body size (in cm) to calculate the body mass index (BMI). For the 354 participants who voluntarily provided information, the mean BMI was 23.7 (SD 3.9) kg/m2.
Recall of the small loss in the mixed-gain outcome conditions.
Recall of the small gain in the mixed-loss outcome condition.
The first hypothesis assumed that for a mixed-gain outcome profile, a close rather than a distant presentation of self-tracking outcomes would lead to an underestimation versus an accurate estimation or overestimation of the recalled small negative outcome. The chi-square test was marginally significant and showed that when mixed-gain outcomes were presented close to (vs distant from) one another, a higher proportion of participants underestimated the small loss (31/106, 29.2% vs 21/113, 18.6%) and a lower proportion of participants exactly recalled or overestimated the small loss (75/106, 70.8% vs 92/113, 81.4%; χ21=3.4,
To test the second hypothesis of how the opportunity to hedonically edit one’s self-tracking outcome affects participants’ health behavior intentions, we examined the mediating role of compensatory health beliefs and the moderating role of health motivation (moderated mediation analysis,
The analysis revealed that health motivation was a marginally significant moderator of the association between hedonic editing opportunity and compensatory health beliefs (estimate of interaction=–0.17;
Results of the moderated mediation analysis (not mean-centered).
Source | Coeff/Effect (SE)a | 95% CI | ||||
Constant | 3.30 (0.32) | 2.67, 3.93 | 10.30 (393) | <.001 | ||
Hedonic editing opportunity | 1.11 (0.49) | 0.14, 2.08 | 2.24 (393) | .03 | ||
Health motivation | –0.04 (0.06) | –0.16, 0.08 | –0.69 (393) | .49 | ||
Hedonic editing opportunity × health motivation | –0.17 (0.09) | –0.35, 0.01 | –1.82 (393) | .07 | ||
Constant | 5.38 (0.37) | 4.66, 6.10 | 14.66 (392) | <.001 | ||
Compensatory health beliefs | –0.21 (0.05) | –0.31, –0.11 | –4.04 (392) | <.001 | ||
Hedonic editing opportunity | –0.28 (0.50) | –1.27, 0.71 | –0.56 (392) | .58 | ||
Health motivation | 0.10 (0.06) | –0.02, 0.22 | 1.59 (392) | .11 | ||
Hedonic editing opportunity × health motivation | 0.02 (0.10) | –0.17, 0.21 | 0.22 (392) | .83 | ||
Low health motivation (mean–1SD=3.94) | –0.09 (0.04) | –0.20, –0.02 | ||||
Health motivation (mean=5.16) | –0.05 (0.03) | –0.12, –0.00 | ||||
High health motivation (mean+1SD=6.38) | –0.00 (0.03) | –0.08, 0.06 | ||||
0.04 (0.02) | 0.00, 0.09 |
aCoefficient for outcome variables and effect for conditional indirect effect.
bOf hedonic editing opportunity on health behavior intentions through compensatory health beliefs at values of the moderator health motivation.
Conditional indirect effect of hedonic editing opportunity on health behavior intentions at values of the moderator health motivation through compensatory health beliefs.
Despite the growing popularity of digital self-tracking apps, to our knowledge this study is the first one exploring how spatial distance in the presentation of multiple self-tracking outcomes (physical activity and diet) affects users’ responses to their self-tracking outcomes. Building on principles of hedonic editing [
Specifically, we found that spatial distance affects users’ attention allocation and resulting memory of self-tracking outcomes consistent with principles of hedonic editing [
Moreover, this research offers a novel explanation of how a hedonic editing opportunity might affect users’ health-related behavioral intentions. Specifically, we provide initial evidence for a motivated cognitive justification mechanism and showed that particularly users’ with a lower health motivation form compensatory health beliefs in response to the opportunity to hedonically edit their outcome, hence leading to reduced health behavior intentions. Thus, for users with a lower health motivation, compensatory health beliefs help to legitimate the hedonic reflection of their self-tracking outcomes and accordingly to justify their reduced intentions for future health-related behaviors in response to their self-tracking outcomes. This finding supports the assumption that “there may be individual differences in the employment of hedonic editing strategies” ([
This research adds to initial work exploring the effectiveness of self-tracking apps to encourage health behavior change [
The limitations of our study offer various avenues for future research. First, we used fictive self-tracking outcomes to create standardized experimental conditions. However, it would be interesting to know whether the observed effects also occur when users view their actual self-tracking outcomes. Second, our study focused on the immediate effects of spatial distance on hedonic editing biases and resulting health behavior intentions and thus cannot provide insights on longitudinal or actual behavioral effects. Accordingly, future research should investigate how spatial distance in the presentation of goal-related outcomes affects health behavior intentions over time and/or how it affects actual physical activity or dietary behavior. Third, we considered self-tracking outcomes from multiple domains that are measured along the same dimension (ie, calories). It would be interesting to examine if multiple self-tracking outcomes that are measured along different measurement units (eg, step count for physical activity and calories burned for dietary behavior) would yield similar hedonic editing effects. Fourth, the participants in our study were not made aware of the possibility that they may experience a hedonic editing bias. To provide insights regarding how the potential negative effect of hedonic editing opportunity on health behavior intentions might be mitigated, it would be worthwhile to investigate whether making users with low health motivation aware of their hedonic editing bias and the potential activation of compensatory health beliefs would attenuate the negative effect of hedonic editing opportunity on health behavior intentions. Fifth, with respect to the manipulation of spatial distance, this study only employed one potential way of presenting goal-related outcomes. In this regard, future research could investigate the integration/segregation of self-tracking outcomes in conditions where spatial distance is even greater (eg, using two separate domain-specific apps) or smaller (eg, one compound chart) than in this study. Finally, the characteristics of our sample limit the generalizability of our research’s findings. Our study employed a convenience online sample and shows how especially younger, majority female, normally weighted, and well-educated users respond to a close compared to distant presentation of mixed self-tracking outcomes and to the opportunity to hedonically edit one’s self-tracking outcomes. Accordingly, further studies might examine whether the observed effects change for older users or users who are less educated or heavily overweight.
Considering that self-tracking apps constitute a promising, cost-effective tool to improve physical activity or dietary behaviors and to promote overall health outcomes [
Our findings suggest that app design features such as spatial distance between the presentation of mixed positive and negative goal-related outcomes can provide users the opportunity to hedonically edit their self-tracking outcome profile (ie, to recall their self-tracking outcomes in a more positive light). As such, if users have a positive outcome in one domain and a smaller negative outcome in the other domain, a spatially close versus distant presentation of self-tracking outcomes could facilitate that users tend to recall the small negative outcome as being smaller than it actually was. Likewise, if users have a large negative outcome in one domain and a small positive outcome in the other domain, a spatially distant versus close presentation of self-tracking outcomes could facilitate that users recall the small positive outcome more accurately. Importantly, particularly among users with lower health motivation, the opportunity to hedonically edit one’s mixed self-tracking outcome profile leads to reduced health behavior intentions. Thus, to improve the effectiveness of self-tracking apps that employ goal-setting techniques for multiple domains, further studies are needed to determine the ideal distance between the presentation of mixed self-tracking outcomes on a digital device’s screen because it is conductive to prevent the occurrence of hedonic editing biases among users and to encourage health behavior intentions.
Details on measures and participant exclusion criteria.
body mass index
This study was completely supported by personal funds.
None declared.