Content uploaded by Yu Chen
Author content
All content in this area was uploaded by Yu Chen on Oct 25, 2016
Content may be subject to copyright.
Exploring Social Accountability for Pervasive Fitness Apps
Yu Chen
Human Computer Interaction Group
Swiss Federal Institute of Technology,
Lausanne, Switzerland
yu.chen@epfl.ch
Jiyong Zhang
Artificial Intelligence Laboratory
Swiss Federal Institute of Technology,
Lausanne, Switzerland
jiyong.zhang@epfl.ch
Pearl Pu
Human Computer Interaction Group
Swiss Federal Institute of Technology,
Lausanne, Switzerland
pearl.pu@epfl.ch
Abstract—Mobile fitness applications have gained increasing
popularity to help users walk and exercise more. A key
component in such apps is its ability to motivate users.
Traditional gamification methods have focused on competition
such as leaderboard for community users, self-reflection for
individual users, or a combination of the two. Motivated by
recent work showing a promising effect of social capital, we
have designed and developed a mobile game, HealthyTogether,
based on such ideas. We are further interested in how users
behave in different settings of gamification methods compared
to a baseline. To this end, we have designed and conducted an
in-depth user study (N=24) involving 12 dyads playing these
games in 4 conditions over a period of two weeks. We report
here the design of the application as well as the user study.
Among the various rewarding schemes, one that uses a hybrid
concept of competition and social accountability gives the most
desirable outcome.
Keywords-health; pervasive fitness applications;
gamification; competition; social accountability.
I. INTRODUCTION
Wellness and lifestyle change have gained significant
attention in recent years. Both research communities and
commercial sectors are putting increasing effort to develop
wearable sensors and mobile applications that help and
“nudge” individuals to increase their physical activities, eat
healthier diet, better manage their sleep and stress, and
engage in social lives with family and friends.
Many of the applications use gamification -- the use of
game elements in non-game context [14] -- to motivate users
to exercise more. Concrete methods include competition
such as leaderboard for community users [9], self-reflection
such as visualization for individual users [2][5][6], or a
combination of the two [7][12]. Recent work shows a
promising effect of social responsibility for the sake of
helping each other especially among family members,
friends, and people who share same interests and goals [1].
We define this concept as social accountability, which refers
to a person’s awareness of another person’s goal and
rendering himself/herself responsible to the goal’s successful
fulfillment.
In this work, we are interested in how users behave in
various settings of gamification methods: competition, social
accountability, a hybrid model of a mixture of competition
and social accountability, and a baseline non-social setting.
To this end, we have developed a mobile application,
HealthyTogether, which enables dyads to participate in
physical activities together, send each other messages, and
earn badges. We use this application as an experimental
platform for an in-depth user study (N=24) to evaluate how
the various reward schemes influence users’ exercises and
social interactions, both quantitatively and qualitatively.
The rest of the paper is organized as follows. After
covering related work in Section II, we present
HealthyTogether in Section III, user study design in Section
IV and results in Section V. We conclude this paper in
Section VI.
II. RELATED WORK
Self-reflection is considered as a self-motivation and
successful strategy for pervasive health applications.
Research prototypes such as Shakra [4] and Houston [13]
and commercial products such as Fitbit and Nike+ all
visualize users’ daily activities to achieve self-reflection. A
number of systems also present physical activities using
metaphors. UbiFit Garden [12] visualizes users’ daily steps
by the growing status of plants. The more activities a user
takes, the healthier his plant looks. Fish’n’Steps [7] uses the
metaphor of fish tank to visualize users’ step count. Recent
work has employed informative art as a visualization tool,
such as research prototype Spark. The above work mainly
motivates users in an individual setting.
Social interaction, including peer-support, cooperation,
competition and belonging to a group has been a clear
motivator for wellness activities [1][8]. Commercial products
have widely adopted competition to motivate user, such as
Nike+ and Fitbit. Fitster [9] is a research prototype that
visualizes users’ steps in a social network and places users in
a virtual competition environment. Kukini [16], Fish’n’Steps
[7] and Life Coaching Application [11] support competition
by helping users to form a team and explicitly introducing
social interaction and social pressure.
Research also shows that social communication can
motivate users to exercise. Consolvo et al. [13] show that
message exchange can help users to increase the
responsibility and give support to group members [9].
Champbell et al. [16] suggest that communication using
everyday fitness games can help enhance players' social
relationship and sustainability in everyday fitness.
Social accountability has been shown to be effective in
helping users to achieve goals. Ahtinen et al. [1] have found
out that connecting with family members and loved ones can
221Copyright (c) IARIA, 2014. ISBN: 978-1-61208-353-7
UBICOMM 2014 : The Eighth International Conference on Mobile Ubiquitous Computing, Systems, Services and Technologies
help motivate users; connecting with people with similar
wellness targets from communities within short distances can
also increase motivation towards wellness activities. Stickk
[15] helps users to achieve their goals by allowing them to
appoint another person to monitor the progress and verify the
accuracy of progress report. They can add supporters who
can encourage them by commenting on their progress. Users
can also put stake on the goal and specify where the stake
would go if they fail in the goal. GoalSponsor [3] allows
users to set up goals and sponsors whom they should be
accounted for. A sponsor can be a friend, a professional in
healthcare, or someone who has accomplished the goal
successfully. Users are more committed in fulfilling the
goals either because they do not want to let others down or
because they do not want to lose reputation in front of others
[3]. In the above work, the structure includes one person who
has a goal to fulfill and another person who monitors the
progress.
To the best of our knowledge, current fitness applications
have not well studied interaction schemes in which users
mutually account for each other’s progress. We are
motivated to investigate social accountability factor in
pervasive fitness applications using an experimental platform
called HealthyTogether.
III. HEALTHYTOGETHER
HealthyTogether is a mobile application that involves a
pair of users to exercise together, and it is implemented on
the Android platform. To measure users' physical activities,
we choose the Fitbit sensor (as shown in Figure 1 a) and b))
among many off-the-shelf sensors as the activity tracker for
our HealthyTogether system. Here, we describe the user
interface design and the underlying rewarding mechanisms.
A. Game Rules
We designed a series of rewarding mechanisms for
HealthyTogether in order to investigate the impacts of
different social settings in pervasive fitness application. A
user can win badges based on Karma Points, which are
calculated as below.
)'()()( ustep sustepsukp ⋅+⋅=
βα
Based on different
α
and
β
values, HealthyTogether
provides the following three reward settings:
• Competition setting, where
α
= 100%,
β
= 0;
• Accountability setting, where
α
= 0,
β
= 100%;
• Hybrid setting, where
α
= 80%,
β
= 20%.
In competition setting, a user's Karma points are calculated
purely by his or her steps. To gain more badges, a user only
needs to focus on his or her own activities even if he is
exercising with a buddy. Thus, we name this rule
competition setting. In accountability setting, a user's
Karma points are calculated by the steps of the buddy.
Therefore, the more he encourages his buddy to exercise,
the more points he earns. Thus, we name it accountability
setting. On the other hand, even if a user does not move at
all, he can still gain badges from the buddy's activities. In
the hybrid setting, a user's Karma points are calculated
based on both his (her) own and that of the buddy,
proportionally. The idea behind this reward scheme is to
encourage competition while also motivating users to cheer
each other. Initially, we set
α
=80%
and
β
=20%
based on
the well-known Pareto Principle. In the future, we will also
experiment different ratios of competition and social
accountability, such as 50%-50% and 20%-80%.
B. Badges
HealthyTogether issues badges based on kp(u). The first
badge is issued if kp(u)>500, to help users get started in a
short time. This number is followed by 1,000 and 2,000 and
then increases by every 2,000 points. HealthyTogether
calculates Karma points in a daily basis but accumulates
badges over time. For example, if a user earns 5,353 Karma
points in a day, he can gain 4 badges, i.e., 500, 1000, 2000
and 4000. If a user earns 5,353 and 6,086 points in the first
two days, he can gain 4 and 5 badges respectively and a total
of 9 badges.
C. Interaction Design
The main interface of the HealthyTogether system is
shown in Figure 1 c). It contains a ‘self’ tab and a ‘buddy’
tab. Each tab displays information about step count, active
time and badges of the current day. We use a pie chart to
visualize the proportion of time that a user is in various
Figure 1. a) The FitBit tracker, b) FitBit in use, and c) the
Samsung Galaxy.
222Copyright (c) IARIA, 2014. ISBN: 978-1-61208-353-7
UBICOMM 2014 : The Eighth International Conference on Mobile Ubiquitous Computing, Systems, Services and Technologies
activity modes, i.e., sitting, lightly active, fairly active and
very active.
The badge area displays the total number and the badges
that a user has earned. The badges are accumulated over
time. In Figure 1 c), the user has earned 6 types of badges
with a total number of 16. When he/she clicks on a badge
icon, a dialog box pops out explaining the details of this
badge type, including how many badges the user has earned
and how he/she earned the badges.
There is a messaging button on the top-right corner of
each page. When it is clicked, users can either view message
history (Figure 2 a)) or send messages to their buddies
(Figure 2 b)). Users will receive a vibrated notification when
buddies send them new messages.
IV. USER STUDY
To study different game settings in real situations, we
designed an exploratory deployment study. We first
conducted a user study (Study 1) that spans for six
continuous working days, which was divided into a three-day
control session and a three-day experimental session. After
conducting the study, we were able to discover some
interesting results. For example, participants suggested that
we extend the study to two weeks, excluding the weekends,
so that the control and the experimental sessions span over
identical days of the week, thus minimizing the influence of
a given day’s schedule to the physical activities being
monitored. For example, a user may work in the office on
Mondays but conduct experiments in the laboratory on
Wednesdays. We therefore conducted the second study (we
name it Study 2) with duration of two weeks. We refer to the
control session as Phase I and experiment session as Phase II
in both studies.
A. Participants
We recruited the participants on campus via word-of-
mouth. After one person signs up, we asked her to invite a
buddy of her choice to join. Their ages range between 22 and
33 and they never used Fitbit before. We required that each
dyad should not work in the same office or too close to each
other. We offered all participants a 50CHF gift card as
compensation for their time.
B. Materials
We provided users with an Android phone with 3G SIM
card and a Fitbit. Three users requested to use their own
Android phones because it would be more convenient for
them. We checked that their phones are compatible for
installing HealthyTogether.
C. Procedure
Both Study 1 and Study 2 were structured as a two-phase,
within-subjects design. Phase I allowed participants to
become accustomed to using Fitbit and allowed us to collect
baseline fitness data. In this phase, all participants use Fitbit
alone without connecting with buddies. In Phase II,
participants in baseline groups (Group A1- A3) continued to
use only the Fitbit while groups in social settings (Group B1-
B3 in competition setting, C1- C3 in accountability setting,
D1- D3 in hybrid setting) started to use Fitbit and
HealthyTogether with buddies. The structure of Study 1 was
the same as the Study 2, except the duration was extended to
two weeks, with both phases extended from three days to
five days.
At the beginning of the study, we invited each pair of
participants to our laboratory and helped them to set up their
Fitbit accounts. We also had a short interview with them on
their experience in using fitness sensors. At the end of Phase
I, we invited participants in social settings to our laboratory
again to install HealthyTogether with different game rules.
Since our user study lasts for up to two weeks, we
requested participants to fill in a daily experience survey
related the study. At the end of each day, we sent a reminder
email with the survey link to participants asking them
whether they have anything to share with us about their
experience using Fitbit or HealthyTogether. The survey only
contains one question: “Do you have anything to share with
us on your experience using Fitbit/HealthyTogether today?”
Daily survey not only helps us to gain an in-depth
understanding of users’ experience, but also facilitates us to
explain their step data with activities during that day.
At the end of the study, we organized a semi-structured
interview. We invited two participants in each group to
attend the session together, so that they could share their
stories. We did not ask a fixed set of questions, but mediated
the session with the following aspects: overall impression,
experience, attitudes and aptitudes, motivation of usage,
social relationship.
V. RESULTS
In this section, we report both quantitative and qualitative
results collected in Study 1 and Study 2. To facilitate
describing results, we encoded the two participants in each
Figure 2. Screenshots of messaging components.
223Copyright (c) IARIA, 2014. ISBN: 978-1-61208-353-7
UBICOMM 2014 : The Eighth International Conference on Mobile Ubiquitous Computing, Systems, Services and Technologies
dyad with ‘a’ and ‘b’ together with their group ID. For
example, we encode the two participants in Group C1 as
‘C1a’ and ‘C1b’ respectively.
A. Quantitative results
Study I
We first investigate users' step count across the 6 days. The
overall average daily step count is 7,439 (min=3,185, max =
11,490). We then compare the average daily step count
between baseline group (A1) and groups who used
HealthyTogether (B1—D1) to evaluate the effectiveness of
social interaction incentives (see Figure 3). Results show a
slight decrease of steps from Phase I to Phase II across all
groups. One explanation is the novelty effect of using Fitbit
in the first 1—2 days, as reflected in daily survey. One
interesting finding is that in Group A1 the average step count
decreased by 20.4% but in Group B1—D1 it decreased by
only 10.6%. This implies that HealthyTogether with social
settings could help users to persist in physical activities.
We further compare Group B1—D1 with different rules
of calculating the accumulative Karma points of the three
days. Results show that users in Group B1 (competition
setting) have largest difference of Karma points (
Δ
kp =
14,556) compared with Group C1 (
Δ
kp = 839) and Group
D1 (
Δ
kp = 2,982). One explanation is that participants in
Group B1 focus more on their own performance compared
with other groups. In other words, it implies that the social
accountability factor, applied in Group C1 and D1, could
lead to more balanced performance of physical activities
between buddies.
Meanwhile, users exchanged 72 messages, shared by
Group B1 (N=43), Group C1 (N=27) and Group D1 (N=2).
The distribution shows that Group C1 (accountability
setting) and D1 (hybrid setting) interact more compared with
Group B1 (competition setting). Particularly, participants in
Group D1 exchanged 59.2% more messages than Group C1.
This implies that hybrid setting is most useful to encourage
participants to interact with each other.
Several topics emerged from the analysis of message
content, and we present them with sample text in Table I.
The distribution of each topic shared in Group B1—D1 is
shown in Figure 4. It reveals the following phenomenon: 1)
in total, there are 27 chat messages, which have the largest
share; 2) encouragement is the main topic that is relevant
with physical activity (20 messages); 3) Group C1 has the
largest share (13 messages) in encouraging messages; 4) the
major topic of Group D1 is workout together (16 messages),
and it only appears in Group D1. The results imply that
hybrid setting introduces most conversation in the topic of
workout together.
Study 2
The deployment of Study 2 is the same as Study 1 except
for the duration. We first verify whether discoveries in Study
1 still exist in Study 2. The average daily step count is 9,501
(min=3,200, max = 24,334). Figure 5 is a distribution of
average daily steps between baseline groups (Group A2, A3)
and social setting groups (Group B2, B3, C2, C3, D2, D3) in
two weeks. The distribution shows a steady increase of
average daily steps in social groups from Phase I to Phase II.
Comparing social setting groups and baseline groups, we
found average steps increased by 9.8% from Phase I to Phase
II in social setting groups but decreased by 10.1% in baseline
groups. This is consistent with implication in Study 1 that
social settings could motivate users to exercise compared to
when they walk alone.
We then compare groups using HealthyTogether in
different social settings. Figure 6 shows each participant’s
average daily steps in Phase I vs. Phase II. The average daily
steps in competition groups (B2 and B3) have increased from
9,747 to 10,128 (
Δ
=381). In accountability groups (C2 and
C3), this number increased from 8,888 to 9,717 (
Δ
=829),
and in hybrid setting (D2 and D3) from 10,762 to 12,437 (
Δ
=1,675). The average daily step increase of hybrid group is
51% more than that of accountability group and three times
more than that of competition group. If we assume that
participants have the same schedule of the same workday in
different weeks, and that Phase I is a baseline for
participants, then the above results suggest that hybrid
setting encourage users to walk more.
Figure 4. Topic distribution of messages exchanged using
HealthyTogether in Study 1.
Figure. 3. Distribution of average daily steps for groups with non-
social setting vs. social setting in Study 1.
TABLE I. MESSAGE TOPICS AND EXAMPLES
Topics
Examples
Self-reflection
“im at 9200.. maybe i can run more”
Cheering
“you should make it 8k for a new badge”
Comparison
“the first time i am higher than you!!! ”
Workout together
“we should walk a round the floor together to
take a break:)”
Chat
“feeling so tired now, go to bed soon”
224Copyright (c) IARIA, 2014. ISBN: 978-1-61208-353-7
UBICOMM 2014 : The Eighth International Conference on Mobile Ubiquitous Computing, Systems, Services and Technologies
We further compare the steps between participants in a
dyad. As shown in Figure 6, dyads from both accountability
groups (C2 and C3) and hybrid groups (D2 and D3) have
increased steps together from Phase II to Phase I. On the
other hand, in competition groups, average step count of B2a
and B3a have increased 22.5% and 11.0% respectively from
Phase I to Phase II while this number decreased for their
buddies (2.1% for B2b and 17.8% for B3b). This implies that
accountability factor helps dyads to walk more steps
together. In Study 1, we found participants in accountability
group and hybrid group are trying to achieve a balanced
number of badges. Even though we do not have the same
finding in Study 2, the results concur with the implication in
Study 1 that users have more balanced working performance
that both users in a dyad improve together.
We then analyze the 86 messages sent between the dyads
in Study 2. Participants in hybrid groups sent 58 messages,
which is more than twice the number of accountability group
(N=21) and seven times more than that of competition group
(N=7). Figure 7 shows the distribution of message topics
within groups of the three social settings. Different from
Study 1, messages with topics about self-reflection and
encouragement have the largest share in the total of
messages (27.8% for both topics). We also discover that
hybrid groups have the largest share of messages (81%) in
the topic of self-reflection. The distribution accords with
what we have found from Study 1 that 1) hybrid groups have
most share of messages (75%) in the topic of workout
together, and 2) encouragement is the major topic (54%) in
accountability groups. If we consider the number of
messages as one metric to evaluate social interaction, results
in Study 2 further provide evidence that hybrid setting is
more likely to stimulate social interactions.
B. Qualitative analysis
In this section, we report the results we found through the
logged daily survey and post-study interviews from both
Study 1 and Study 2.
During the user interview, we found some qualitative
results that are related to our findings from quantitative
analysis. Overall, the feedback about HealthyTogether was
very positive. First, HealthyTogether has helped them to
compare with each other. As B1b said in the interview, “to
check her (buddy's) steps and compare with mine is most
important for me”. Second, they could interact with each
other via HealthyTogether. B1a reported in survey: “I
received message on the first day, so the next day I
intentionally walked more between the buildings.”
We also found some evidence that the accountability
factor (applied in accountability and hybrid setting) could
help users to care about each other. As C1b reported in her
daily survey, “I discovered his step is twice more than mine.
As his badges depend on my steps, I feel I should walk more
in order not to discourage him.” This supported what we
found in Study 1 that the participants in C1 have more
balanced performance when using HealthyTogether.
Additionally, when we asked whether users about their social
relationship before and after using HealthyTogether,
participants in D3 revealed that they were already very close
friends. Participants in C1, C2, and D1, and D2 reported that
they had developed further relationship with buddies after
using HealthyTogether. For example, D1b said: “Even
though we are colleagues, we did not talk much. Finding
something to do together rapidly brought us closer.” For
another example, C2a reported that she knows more about
her buddy: “When I woke up, my buddy already had 3,000+
steps... She is already on campus...” However, we did not
find the same report from competition groups. This suggests
that social accountability could be helpful for a user to
enhance social relationship with the buddy.
Participants have reported their concerns regarding
competition setting. Both B1a and B1b reported competing
with each other cause demotivation. “I knew I would never
beat him because he needs to walk a lot from home to
Figure 5. Distribution of average steps non-social setting vs. social
setting in Study 2.
Figure 6. Average step count Phase I vs. Phase II in Study 2.
Figure 7. Topic distribution of messages in Study 2
225Copyright (c) IARIA, 2014. ISBN: 978-1-61208-353-7
UBICOMM 2014 : The Eighth International Conference on Mobile Ubiquitous Computing, Systems, Services and Technologies
work…” –B1b. “I clearly had advantage in winning the
game and sometimes I’m afraid walking too much makes her
pressure…” –B1a. B3a also reported that when he noticed
his buddy walking less, he also became lazier. Admittedly,
the demotivation effect could be caused by the unbalanced
abilities in physical exercise. This suggests that choosing a
suitable exercise buddy can be important in order to
maximize the effect of competition.
Evidence shows that the hybrid setting is more preferred
than the accountability setting. “I feel a little bit wired if my
badges only depend on my buddy's steps. Is it a little bit de-
motivating me?” mentioned by C3a. By comparison, instead
of ‘depending on’ others, both D1 and D3 have reported they
arranged activities together, such as walking to a farther
away cafeteria to have lunch (D1), and going to Zumba
course together (D3). D2a reported in his survey that his
buddy gave him suggestions to increase the steps: “Without
trying too hard, I almost reached 10k steps. Seeing his
progress during the day motivated me to move more. It was
also useful to talk to him (via message) - he gets a high step
count by walking around while reading articles. I also did
this walking in circles thing for about 1h, which helped my
brainstorming.” Even though D2b (avg=20,372) have clear
advantages over D2a (avg=5,852) in Phase I, this
discrepancy did not demotivate either of them. Instead,
D2a’s average daily step has increased by 57.3%
(avg=9,207) from Phase I to Phase II and D2b has increased
by 11.2% (avg=22,661). As D1a said: “Helping others to
become better is a `plus' rather than `minus' to your life.”
VI. CONCLUSIONS
In this work, we have developed a mobile application
called HealthyTogether that allows dyads to participate in
daily physical exercises as a game. We conducted an in-
depth user study with 12 dyads with a period of up to 2
weeks and compared participants using HealthyTogether in 3
social settings and a baseline non-social setting. Results
show that social settings, even in the competition mode, can
help users to persist more in physical activities compared
with baseline group. Additionally, the hybrid setting is more
likely to motivate users to walk more and more actively help
others. Furthermore, the number of messages sent between
participants in the hybrid settings is 8 times more than those
in the competition setting and twice of those in the
accountability setting. Integrating social accountability factor
is also promising to enhance social relationship between
buddies.
In the future, we plan to conduct longitudinal studies
with more users in various conditions to validate our findings
with statistical analysis.
ACKNOWLEDGEMENT
We thank Swiss National Science Foundation for
sponsoring this research work. We are also grateful for all
anonymous reviewers for their valuable feedback.
REFERENCES
[1] A. Ahtinen, M. Isomursu, M. Mukhtar, J. Mäntyjärvi, J.
Häkkilä, and J. Blom. Designing social features for mobile
and ubiquitous wellness applications. The 8th international
Conference on Mobile and Ubiquitous Multimedia
(MUM’09), ACM. Nov. 2009.
[2] C. Fan, J. Forlizzi, and A. K. Dey. Spark Of Activity:
Exploring Informative Art As Visualization For Physical
Activity. Proc. of the ACM Conference on Ubiquitous
Computing (Ubicomp’12), ACM. Sept. 2012, pp. 2–5.
[3] GoalSponsor. https://www.goalsponsors.com. Accessed:
August 6, 2014.
[4] I. Anderson, J. Maitland, S. Sherwood, L. Barkhuus, M.
Chalmers, M. Hall, and H. Muller. Shakra: tracking and
sharing daily activity levels with unaugmented mobile
phones. Mobile Networks and Applications, 12(2-3), 2007,
pp. 185-199.
[5] I. Li, A. K. Dey, and J. Forlizzi. A stage-based model of
personal informatics systems. Proc. of the SIGCHI conference
on Human Factors in computing systems (CHI’10), ACM.
Apr.2010, pp. 557-566.
[6] I. Li, A. K. Dey, and J. Forlizzi. Understanding my data,
myself: supporting self-reflection with ubicomp technologies.
Proc. of the ACM Conference on Ubiquitous Computing
(Ubicomp’11), ACM. Sept. 2011, pp. 405-414.
[7] J. Lin, L. Mamykina, S. Lindtner, G. Delajoux, and H. Strub.
Fish’n’Steps: Encouraging physical activity with an
interactive computer game. UbiComp 2006, pp. 261-278.
[8] J. Maitland, and M. Chalmers. Designing for peer
involvement in weight management. Proc. of the SIGCHI
conference on Human Factors in computing systems
(CHI’09), ACM Press (2011), pp. 315-324.
[9] N. Ali-Hasan, D. Gavales, A. Peterson, and M. Raw. Fitster:
social fitness information visualizer. In CHI'06 Extended
Abstracts on Human Factors in Computing Systems. Apr.
2006, pp. 1795-1800.
[10] P. Klasnja, S. Consolvo, and W. Pratt. How to evaluate
technologies for health behavior change in HCI research.
Proc. of the SIGCHI conference on Human Factors in
computing systems (CHI’11), ACM. May 2011, pp. 3063-
3072.
[11] R. Gasser, D. Brodbeck, M. Degen, J. Luthiger, R. Wyss, and
S. Reichlin. Persuasiveness of a mobile lifestyle coaching
application using social facilitation. Persuasive Technology
2006, pp. 27-38.
[12] S. Consolvo, D. W. McDonald, T. Toscos, M. Y. Chen, J.
Froehlich, B. Harrison, and J. A. Landay. Activity sensing in
the wild: a field trial of UbiFit garden. Proc. of the SIGCHI
conference on the SIGCHI conference on Human Factors in
computing systems (CHI’08), ACM. Apr. 2008, pp. 1797-
1806.
[13] S. Consolvo, K. Everitt, I. Smith, and J. A. Landay. Design
requirements for technologies that encourage physical
activity. Proc. of the SIGCHI conference on Human Factors
in computing systems (CHI’06), ACM. Apr. 2006, pp. 457-
466.
[14] S. Deterding, M. Sicart, L. Nacke, K. O'Hara, and D. Dixon.
Gamification: using game-design elements in non-gaming
contexts. In CHI'11 Extended Abstracts on Human Factors in
Computing Systems, ACM. May 2011, pp. 2425-2428.
[15] Stickk. http://www.stickk.com. Accessed: August 6, 2014. 2
[16] T. Campbell, B. Ngo, and J. Fogarty. Game design principles
in everyday fitness applications. Proc. of the 2008 ACM
conference on Computer supported cooperative work
(CSCW’08), ACM. Nov. 2008, pp. 249-252.
226Copyright (c) IARIA, 2014. ISBN: 978-1-61208-353-7
UBICOMM 2014 : The Eighth International Conference on Mobile Ubiquitous Computing, Systems, Services and Technologies
