=Paper=
{{Paper
|id=Vol-3669/paper17
|storemode=property
|title=Promoting green mobility through gamified transportation campaigns
|pdfUrl=https://ceur-ws.org/Vol-3669/paper17.pdf
|volume=Vol-3669
|authors=Simone Bassanelli,Federica Gini,Antonio Bucchiarone,Annapaola Marconi
|dblpUrl=https://dblp.org/rec/conf/gamifin/BassanelliGBM24
}}
==Promoting green mobility through gamified transportation campaigns==
https://ceur-ws.org/Vol-3669/paper17.pdf
Promoting green mobility through gamified
transportation campaigns
Simone Bassanelli1,2, Federica Gini1,2, Antonio Bucchiarone1 and Annapaola Marconi1
1 Fondazione Bruno Kessler, Via Sommarive 18, Trento, 38123, Italy
2 University of Trento, Address, Via Calepina 14, Trento, 38122, Italy
Abstract
This paper explores the impact of gamified campaigns within the AirBreak project on promoting sustainable
urban mobility. Designed to induce lasting behavior change, the campaigns target diverse scenarios, including
home-to-work, home-to-school, and leisure travel. The study assesses the overall effectiveness of these
campaigns, analyzing changes in individual mobility choices and their contribution on reducing CO 2 emissions.
Additionally, the paper delves into the motivational aspects of the campaigns, identifying features that
influenced participants towards sustainable mobility. Understanding these motivational factors provides
insights into key elements driving positive behavioral changes. Moreover, the research investigates user
characteristics influencing consistent engagement in sustainable mobility choices. This analysis enhances our
understanding of the factors shaping participants’ inclination towards adopting and sustaining eco-friendly
travel practices. Grounded in theoretical foundations, the study details data collection and analysis methods,
presenting findings that offer valuable perspectives for future interventions and policy considerations in the
context of gamified sustainable transportation campaigns.
Keywords
gamification, green mobility, green transportation, behavior change, sustainability1
initiatives aimed at increasing citizens’ awareness and
1. Introduction involvement in the transformative process, influencing
their mobility habits in a gradual yet profound manner
Mobility assumes a crucial role in contemporary urban [1].
settings, shaping the way citizens engage with the city, In many instances, citizens’ everyday mobility
access essential services, and participate in urban life. decisions are shaped by ingrained habits and
The effectiveness and organization of mobility, as influenced by inaccurate or outdated beliefs [5]. It is
highlighted by Vesco et al. [1], significantly influence essential for citizens to be well-informed about the
citizens’ experiences within the city. In this dynamic mobility services available in their city and their actual
context, cities grapple with a formidable challenge. value, encompassing considerations such as time, cost,
Administrators must not only ensure citizens’ right to and environmental impact. A heightened awareness is
mobility and seamless access to local services, but also necessary for individuals to recognize the
strive to minimize the economic, social, and repercussions of their daily choices, including their
environmental costs associated with the mobility impact on traffic, greenhouse gas emissions, and social
system. Addressing this challenge necessitates a costs [6]. Crucially, citizens should perceive
comprehensive approach that efficiently leverages themselves as integral to a community where their
existing mobility resources while integrating and collective daily choices play a pivotal role in advancing
promoting new or emerging mobility services, city-level mobility strategic objectives [7]. In essence,
fostering an integrated, efficient, and sustainable individuals and communities must cultivate a sense of
mobility ecosystem [2]. In pursuit of this objective, responsibility, contributing to the development of a
cities are strategically planning and implementing new cultural paradigm for both urban and rural
interventions across infrastructures, services, and mobility.
mobility policies. While these elements are pivotal in In recent years, substantial efforts have focused on
advancing sustainability and integration in mobility, it utilizing interactive technologies to enhance citizen
is crucial to recognize the equally significant socio- awareness, promote active participation, and induce
technical dimension of user acceptance and adoption behavior change toward a more sustainable lifestyle.
[3, 4]. Innovative policies, infrastructures, and services Gamification, identified as a persuasive technology
run the risk of falling short if not complemented by with significant potential [8, 9], is increasingly
8th International GamiFIN Conference 2024 (GamiFIN 2024), April 2-
5, 2024, Ruka, Finland.
sbassanelli@fbk.eu (S. Bassanelli); federica.gini@unitn.it (F. Gini);
bucchiarone@fbk.eu (A. Bucchiarone); marconi@fbk.eu (A. Marconi)
0000-0001-6061-8169 (S. Bassanelli); 0000-0003-3427-3747 (F.
Gini); 0000-0003-1154-1382 (A. Bucchiarone); 0000-0001-8699-
7777 (A. Marconi)
© 2024 Copyright for this paper by its authors. The use permitted under
Creative Commons License Attribution 4.0 International (CC BY 4.0).
CEUR
ceur-ws.org
Workshop ISSN 1613-0073
Proceedings
183
recognized for its applicability in the mobility domain personalized cooperative and competitive game
[10, 11, 12, 13, 14, 15], as well as in various other mechanics. The goal is to encourage and sustain the
environmental sustainability domains [16, 17, 18, 19]. adoption of more sustainable mobility habits among
The core concept involves harnessing the motivational the populace. Moreover, AIR-BREAK focuses on
and persuasive influence of games through the design community building. By actively involving citizens in
of systems that effectively utilize and integrate game AIR-BREAK initiatives and events, the project aims to
concepts and elements [20, 21]. Gamified systems foster the creation of a local community of users.
incentivize individuals to make specific decisions or Tailored behavioral change programs are deployed to
perform essential tasks, crucial for achieving valuable reach diverse user segments, including primary,
objectives [22]. This transformation occurs by secondary, and high school students, as well as
converting potentially unattractive actions into employees. The intention is to create synergies with
enjoyable and rewarding experiences [21]. other AIR-BREAK initiatives, promoting a
The development of urban mobility policies has collaborative and supportive environment. Finally,
become crucial for governments and stakeholders AIR-BREAK emphasizes the importance of producing
aiming to support sustainability goals [23]. Research measurable outcomes. This entails evaluating the
indicates that active and sustainable mobility can play project’s impact in terms of engagement, retention,
a vital role in reducing greenhouse gas emissions [24]. increased awareness, and behavioral change. The
Despite the societal and individual advantages of insights gained from these measurable outcomes will
active mobility, there is a challenge in convincing be instrumental in refining and optimizing subsequent
people to shift from car-centric lifestyles [25]. phases of the AIR-BREAK project, ensuring continuous
Therefore, we implemented two gamified sustainable improvement based on real-world experiences.
mobility campaigns providing more incentives for Within the framework of the AIR-BREAK project,
users to adopt environmentally sustainable methods we have conceptualized and executed diverse
of travel. sustainable mobility campaigns focusing on distinct
The two campaigns, incorporated within the AIR- mobility contexts, including home-to-work mobility,
BREAK project2, have been designed and implemented home-to-school mobility, leisure, and free time mobility.
across various contexts such as commuting between These initiatives cater to a range of end-user groups,
home and work, home-to-school travel, leisure encompassing the general public, students, and
activities, and free time mobility. The aim of this employees. The subsequent section provides a concise
initiative is not only to alter users’ mobility patterns overview of each of these planned actions. AIR-BREAK
during the campaign but also to instill a lasting mobility campaigns have been implemented through
behavior change. the Play&Go platform that supports the definition and
In this paper, we present an exhaustive description management of different types of sustainable mobility
of the AIR-BREAK project and the implemented campaigns, customizable with respect to mobility
campaigns (Section 2). In Section 3, we present the objectives and target users. The customization of
theory and hypotheses that guide this paper. In Section campaigns can concern: the means of mobility to be
4, we introduce the methods used for the promoted, the travel validation criteria, the
questionnaire creation and evaluation, data collection, competitive/cooperative game elements
and statistical analysis. Then, in Sections 5 and 6, we implemented, and the real incentives provided by the
report and discuss the findings after the analysis. We campaign [26]. Users interact with the system through
conclude the paper with the conclusions and future the Play&Go App (see Figure 1), which supports
works related to the AIR-BREAK project in Section 7. various functionalities. In addition to player
registration, discovering and registering for active
campaigns, managing the player’s profile, and
2. The AIR-BREAK Project inspecting campaign results, the app incorporates a
robust tracking system able to track single and
The AIR-BREAK project is centered on the core multimodal sustainable trips. This system collects
objective of informing and heightening citizens’ journey data to assign rewards, modify the game status
awareness about sustainable mobility services, with for players based on their activities, and calculate the
the goal of fostering the adoption of eco-friendly travel reduction in CO2 emissions associated with
habits. In outlining the project’s scope, various sustainable transportation choices instead of car use.
campaigns have been designed to address four key The app employs a specific formula that considers the
objectives. Firstly, the project endeavors to actively kilometers traveled using different sustainable
engage citizens, making them aware of both existing transportation means available in various mobility
and newly introduced mobility resources. campaigns. Furthermore, users can access information
Simultaneously, efforts are directed towards elevating on weekly and global rewards, as well as the rules and
their awareness regarding the significant impact their regulations of different campaigns and engage in filling
daily mobility choices can have. The ultimate ambition out questionnaires. This multifaceted approach
is to contribute to the establishment of a new cultural enhances user engagement, promotes sustainable
norm for urban and rural mobility within the city mobility practices, and contributes to both individual
community. Building on this foundation, AIR-BREAK rewards and environmental conservation efforts.
seeks to instigate a voluntary travel behavior change.
This is achieved through a strategic blend of virtual
(game-based) and tangible incentives, leveraging
2 https://airbreakferrara.net/
184
Figure 1: The Play&Go App.
weekly challenges. This personalization is achieved
through a Recommendation System, ensuring that
2.1. Urban Mobility Campaign challenges are not only motivating but also realistic,
avoiding frustration stemming from unattainable
The Urban Mobility Campaign (UMC) is an initiative targets [30]. The game seamlessly integrates both
encouraging citizens to actively embrace sustainable single-player and multiplayer challenges, catering to a
transportation through gamification. Registration and diverse range of player preferences. Single-player
utilization of the Play&Go app for tracking eco-friendly challenges span various types, including performance-
trips are essential components of engagement. Central based, repetitive behavior, surveys, and events, each
to the gaming experience in the UMC are the Eco- contributing to the player’s Eco-Leaves points. In
Leaves Points. These points serve to keep the user multiplayer challenges, players experience a sense of
motivated [27] and as a cornerstone for progression in community and relatedness. These challenges come in
both weekly and global leaderboards [28], thereby cooperative, time-based competitive, and
influencing the allocation of weekly and final prizes. performance-based competitive modes. Cooperative
Earned for each tracked and validated sustainable trip, challenges necessitate cumulative effort from
the calculation hinges on factors such as the distance participants, fostering a collaborative spirit.
covered and the sustainability of the chosen means of Meanwhile, competitive challenges introduce a
transportation. The core mechanics of the game captivating dimension with fixed targets, time
revolve around trip tracking, wherein players specify constraints, and performance-based criteria.
their mode of — be it walking, biking, public A pivotal aspect of the UMC is the continuous
transportation, or car-pooling. An automatic evaluation of player performance through game levels.
validation process ensures the accuracy of each trip, Players progress through different levels such as Green
avoiding potential game abuse [26]. Lover, Green Warrior, and Green Guru based on the
A thoughtful design choice involves the total points they accumulate. Importantly, this system
introduction of daily limits on tracked kilometers and ensures that certain features of the game become
trips to prevent disproportionate and unnecessary use available to players upon reaching specific levels,
of transportation means. This limitation ensures a adding an element of progression and achievement to
more equitable gaming experience for all participants, the gaming experience. The gameplay adopts a weekly
narrowing the performance gap between top players structure, fostering a sense of regularity and allowing
and others actively engaged in the campaign. for the creation of weekly leaderboards. These
Furthermore, constraints tied to the validity of tracked leaderboards, reflecting performance within a specific
trips based on location — valid only if the origin or time-frame, contribute to the excitement of weekly
destination falls within the specific territory — add an prizes offered by local sponsors. Importantly, this
additional layer of relevance and authenticity to the structure provides newcomers with a fair opportunity
gameplay. Beyond the core mechanics, players have to compete with seasoned players, leveling the playing
opportunities to earn bonus Eco-Leaves points. These field and maintaining a vibrant gaming community. To
incentives include inviting friends to join the game, incentivize sustained participation, the UMC concludes
achieving success in weekly challenges, and obtaining with final prizes awarded to top players on the global
special badges that come with associated Eco-Leaves leaderboard. Both weekly and final prizes are
bonuses. Participants can meticulously monitor their generously offered by local associations and sponsors,
mobility history and achievements through enhancing the overall appeal of the campaign.
personalized profiles, which showcase a diverse array
of badges symbolizing accomplishments. These badges
range from achieving specific Eco-Leaves milestones 2.2. High School Challenge
to showcasing preferences in transportation modes,
reinforcing the exploration of various mobility The High School Challenge (HSC) is a competition
alternatives [29]. designed for high school classes, with the goal of
Distinguishing itself from a one-size-fits-all encouraging sustainable and active commuting from
approach, the UMC thrives on its highly personalized home to school. Participants utilize the Play&Go
185
application to track their eco-friendly journeys. Each The first research question stems from the fact that
student not only competes individually in the UMC but the use of a gameful system does not mean guaranteed
also contributes to their class’s overall standing in the success [20, 33]. Then, it is crucial to report whether
inter-school competition. Attractive prizes await the the campaign was effective in enhancing participants’
most active and sustainable classes. The class sustainable choices in terms of mobility. The second
registration process involves the following steps, research question stems from the need to identify
facilitated by a designated reference teacher: whether specific elements could be useful in the
1. Participants download the Play&Go creation of future sustainable mobility campaigns. The
application and register for the game. last research question is based on data in the literature
2. The reference teacher, using an institutional [34, 35, 36] suggesting that users’ demographic
email account, communicates the group data characteristics can modulate their appreciation for
through the registration form. gameful systems.
3. Teams are formed to compete in the
competition’s rankings.
Teams must comprise students from the same 4. Methods
class, with a minimum of 10 students per team or the
entire class for classes with fewer than 10 students. 4.1. Questionnaire development
Classes achieving at least 90% participation receive an
initial bonus of 300 Eco-Leaves points. Teachers Two questionnaires were used to evaluate users’
interested in joining can also be part of the class team, appreciation for the campaign and behavioral change
with a limit of one team per teacher. Participants in terms of sustainable mobility. The first pre-
continue to use the Play&Go application to track their campaign questionnaire consists of six categorical
sustainable journeys. Each participant not only items related to the background information on users’
competes individually but also contributes to their preferences for means of transportation — including
team’s performance in the inter-school competition. In bus, train, bike, walk, and car —, and one categorical
this competition, all Eco-Leaves points earned by team item for information about the app. The post-campaign
members throughout the competition’s duration using questionnaire repeats the same questions as the first
the Play&Go App are considered. The team one, and then, for the core part, a model with five latent
accumulating the highest number of Eco-Leaves and 13 manifest variables was formulated following
points, calculated as the average value of its members, the results of previous campaign analyses [31, 32].
emerges as the winner. These items were collected using a 5-point Likert-scale
Throughout the competition, the global with a magnitude from 1 (Disagree) to 5 (Agree).
leaderboard of classes can be accessed via the Play&Go Lastly, three open questions collected detailed
App and the AIR-BREAK website. Upon the information on appreciated and non-appreciated
competition’s conclusion, the teams leading the HSC elements of the campaign3. The theory of the core part
Global Eco-Leaves Points Rankings will be awarded consisted of five core constructs: intrinsic motivation
collective prizes. (IM), extrinsic motivation (EM), behavior change (BC),
tool attractiveness (TA), and future behavior (FB), and
a single item for the evaluation of the Overall
3. Theory and hypotheses appreciation (OA). Moreover, items related to the
appreciation of game elements (TA2 — [Specific game
Given the various gamified campaigns in the AIR- element] made it enjoyable to participate in the
BREAK project, all aligned with a common theme, we [campaign] initiative), and those related to behavior
identified the need to develop a cross-sectional change due to game elements (BC3 — [Specific game
questionnaire applicable across all campaigns that element] prompted me to go more often by
incorporate game elements. From a long-term analysis environmentally friendly means) were repeated for
of previous campaigns [31, 32], we identified a each game element entered into the campaign to
different interaction between factors in relation to the assess how much specific game elements contributed
nature of different rewards and game elements, and to campaign appreciation and actual behavioral
specific correlations between some factors, such as change.
campaign appreciation and behavioral change.
Therefore, we decided to base the new questionnaire
on specific relationships. Then, we formulated three 4.2. UMC participation
research questions.
RQ1. To what extent was the campaign effective in The UMC (introduced in Section 2.1), ran from April
enhancing participants’ sustainable choices in terms of 17, 2023, to September 24, 2023. Ferrara citizens
mobility? tracked their sustainable movements through the
RQ2. Is the preference for specific features of the Play&Go app, including cycling, walking, taking the
campaign related to users’ motivation in making bus, train, or carpooling. Participants actively took
sustainable choices in terms of mobility? part in personalized mobility challenges, striving to
RQ3.Are there any characteristics of the users that climb the rankings and win various weekly and final
make them more prone to appreciate the campaign? prizes. The results of the campaign have been highly
3 The full questionnaire can be retrieved here:
https://osf.io/6mrga/?view_only=d9ecfb3b679e41ecbb0fde89404b
2e7b
186
encouraging in terms of participation. As depicted in analysis. SEM investigates the connections among
Figure 2, a total of 258 citizens joined the initiative, variables, encompassing both observable and latent
with 55% being women and 45% men. In terms of age ones, within a model specified by the researcher based
distribution, participants between the ages of 35-50 on theoretical considerations and prior findings. The
accounted for 37.8%, while those in the 20-35 age model parameters are typically estimated using
group constituted 17.07%. Individuals between the techniques like maximum likelihood from the
ages of 50-70 comprised 43.9%, and those under 20 covariance matrix of the observable variables [37, 38].
made up 1.22%. To provide us with detailed feedback Statistical evaluation of model fit is often performed
on their enjoyment and effectiveness in behavioral using the chi-square (𝜒²) goodness-of-fit test.
change, the questionnaire was sent to participants at Additionally, the Root Mean Square Error of
the end of the campaign. Approximation (RMSEA), derived from the 𝜒² test,
provides a measure on an absolute scale, reflecting the
model’s fit to the data while considering factors such
as model size and sample size. In this study, RMSEA
values below 0.05, Comparative Fit Index (CFI) values
above 0.96, and Standardized Root Mean Square
Residual (SRMR) below 0.08 are considered indicative
Figure 2: UMC Participation and Impact.
of a satisfactory fit [37, 39, 40], indicating that there is
no need to arrange items and the structure differently.
4.3. HSC participation Lastly, many authors do not rely on the analysis of
𝜒² value and its significance, as it is highly subject to
The HSC (introduced in Section 2.2) lasted from April sample size [41]. We therefore preferred to rely on the
3 to June 4, 2023, and utilized the Play&Go App. ratio of chi-square to degrees of freedom (df). Values
Through this application, students and teachers less than 2 are often considered acceptable. MPLUS
tracked their sustainable movements (cycling, (version 8.4)4 was used to run the CFA.
walking, bus, train, or carpooling), engaging in team
play, and climbing the rankings to win final prizes. 4.5. Questionnaire analysis
4.5.1. Participants
We examined information gathered from 117
participants (female = 67, aged between 16 to 70 years
Figure 3: HSC Participation and Impact. old), who actively participated in two campaigns: HSC
(N = 26) and UMC (N = 91), and completed the final
Notably, the HSC featured 7 teams (see Figure 3), questionnaire through Google Form. Of the 117
with 5 teams composed of students and 2 teams participants who completed the final questionnaire, 74
consisting of teachers, adding a collaborative (all from the UMC campaign) had also completed a pre-
dimension to the challenge. The overarching goal of campaign questionnaire.
the HSC was to promote sustainable commuting
between home and school, as well as during leisure 4.5.2. Analysis
time, for all students aged 14 and above. This objective
was achieved through cooperative efforts within
We proposed an analysis method for the
classes and inter-class competition among high
questionnaire. Not to weight the number of items
schools. To provide us with detailed feedback on their
related to the number of game elements, any item
enjoyment and effectiveness in behavioral change, the
related to the appreciation of game elements was
questionnaire was sent to participants at the end of the
averaged (TA2), as were those related to behavior
campaign.
change due to game elements (BC3). In addition, as
defined in other scales, such as MEEGA+ [42], and SUS
4.4. Confirmatory factor analysis [43], we identified a unique final value to provide a
holistic evaluation of the
Before analyzing the data comprehensively, we campaign, defined as “total score”, ranging from 0 to 1,
carefully checked the theoretical arrangement of the based on the formula 𝑋⁄5, where 𝑋 represents the
items against the data obtained through a mean of the answers calculated for each user. The
confirmatory analysis, and then grouped the items choice of this formula falls back on the fact that, thanks
differently within the questionnaire. Confirmatory to the 5-point Likert-type answers, it is possible to
factor analysis (CFA) is employed to assess and obtain a score ranging from 0 to 1, making it feasible to
potentially refine a predefined model depicting the create a priori thresholds for campaign satisfaction,
relationships between latent variables (factors) and and because a successful sustainable mobility
observable (or measured) variables. CFA falls under campaign assumes that game elements and the
the umbrella of Structural Equation Modeling (SEM), a platform have been implemented properly to support
statistical method that amalgamates multivariate behavioral change during the campaign, and in the
techniques such as regression analysis and factor long term. Furthermore, these elements succeed in
4 https://www.statmodel.com/
187
motivating users who are driven by intrinsic or the post-campaign survey and in particular, analyzing
extrinsic motivation equally. The choice of using a total items related to behavior change (items BC1 and BC2),
comprehensive score stems from the need for the results showcased the capacity to change players’
unambiguous value to understand how successful behaviors: 94% of players stated they felt motivated to
sustainable mobility campaigns have been perceived. adopt more sustainable mobility habits. Specifically,
The questionnaire’s design allows for the evaluation of 79% incorporated sustainable mobility habits into
how various game elements impact campaign their daily commutes, and 77% for leisure travel.
appreciation and user motivation. Additionally, it Finally, a significant outcome is the participants’
facilitates the summary, on one hand, and comparison, satisfaction, with 98% expressing a definite
on the other, of the distinct campaigns within the AIR- willingness to participate in future editions (item FB2).
BREAK projects. While a more in-depth analysis of This initiative also led to the formation of a local
each individual campaign is planned for the future, this community of motivated and active users advocating
paper concentrates on the overarching aspects shared for more sustainable mobility.
by both campaigns. To answer to RQ1, we evaluated
the effectiveness of the campaign by comparing
participants’ sustainability scores at the beginning and
at the end of the campaign through a Wilcoxon test for
paired samples. Sustainability scores were calculated
using the following formula:
(𝑏𝑢𝑠 + 𝑡𝑟𝑎𝑖𝑛 + 𝑏𝑖𝑘𝑒 + 𝑤𝑎𝑙𝑘)⁄4 + (7 − 𝑐𝑎𝑟)
2
with bus, train, bike, walk, and car being participants’
answers to a 6-point Likert scale on their mobility
habits (1 = almost never, 6 = more than once a day).
We examined participants who completed both the
initial and final questionnaires to assess the
effectiveness of the platform and gain insights into
how users’ habits before using the app might impact Figure 4: Use of different transportation means in
post-results. We ran a series of Spearman correlations UMC.
between EM, IM, the total score and (i) pre-
sustainability scores, (ii) post-sustainability scores, For what concerns the HSC campaign, more than
and (iii) ∆ scores, representing the difference between 100 students and teachers have participated in the
pre and post campaign scores. Furthermore, a HSC initiative, with over 3,000 tracked journeys, in
Wilcoxon test was used to compare participants’ Play&Go, covering more than 9,000 sustainable
scores in the EM and IM constructs. To answer to RQ2, kilometers (of which 27% were by bike and another
we analyzed participants’ open-ended responses to 27% walking - zero impact), and nearly 2 tons of CO2
the question "What did you like the most about the saved (see Figure 5). Analyzing the post-campaign
campaign?" we categorized them into two groups survey, the experiment has demonstrated the ability to
based on their answers. Users’ answers were short and change players’ behaviors (items BC1 and BC2):
referred to specific elements of the campaign (e.g., almost 70% of players claim to have adopted more
"rewards", "points", "the initiative", "the idea"). We sustainable mobility habits in their home-school (or
used a Mann-Whitney test to compare participants’ home-work) commutes, while 75% have improved
total scores. Then, we ran a rank-transformed ANOVA their habits during leisure time. Finally, an important
to compare IM and EM scores based on the elements result (item FB2) is the satisfaction of the participants
users valued most. Finally, to answer to RQ3 a rank- (70% express a desire to participate in future editions)
transformed ANOVA was used to explore whether and the creation of a local community of motivated and
demographic characteristics, such as age and gender, active users for more sustainable mobility.
influenced users’ total scores.
5. Results
5.1. Participation results
During the UMC campaign, exploiting the Play&Go app,
nearly 31,000 journeys have been recorded that
covered over 148,000 sustainable kilometers, with
over 95,000 by bike and almost 15,000 walking
(resulting in zero impact) (see Figure 4). The
experiment demonstrated the ability to sustain
citizens’ motivation in long-term campaigns, with
continuous and consistent participation over the Figure 5: Use of different transportation means in
initiative’s duration of more than 5 months. Exploiting HSC.
188
5.2. Questionnaire results appreciation for the project (N = 117, M = 0.797, SD =
0.147). As is often the case with questionnaire data
The test statistics for this model were RMSEA = 0.039, with Likert-type scales [44], no data were normally
𝜒²/df = 1.176, CFI = 0.986, and SRMR = 0.050. These distributed within the measured variables. First of all,
results indicated that the model fit was good (Figure we were interested in understanding the increase in
6). The data also indicated significant relationships participants’ sustainability choices in terms of
between several latent factors, however, two transportation, and how their old habits are related to
hypothesized relationships (IM-BC, and EM-TA) were their final overall score, along with the five constructs
not significant, so they were not reported in the model (RQ1). To achieve this, we included in the analysis
structure. See Table 1 for a comprehensive overview. those participants who completed both the pre- and
The results show that users’ behavior change can be post-campaign evaluations (N = 74, due to missing
explained by extrinsic motivation, thus conveyed answers in the pre-campaign evaluation. All 74
primarily by the rewards, while tool and game participants are from the UCM campaign). Given the
elements’ attractiveness seem to be explained by non-normal distribution of the data, non-parametric
intrinsic motivation. Interestingly, the future behavior tests were employed for the analysis (refer to the
is explained by both users’ expressed behavior change supplementary material for normality analyses6). We
values and tool attractiveness. As hypothesized, tool utilized a Wilcoxon test for paired samples to compare
appreciation and behavioral change reports go hand in sustainability scores based on participants’ habits in
hand. Lastly, the final score for the campaign’s overall terms of sustainable mobility before and after using
appreciation, related to the item "Participating in the application. The results revealed a significant
[campaign] was enjoyable overall", is explained by the difference in participants’ sustainable habits before
interaction with tool attractiveness and future and after using the application (W = 371.5, p-value <
behavior. Based on these data, we were able to proceed 0.001). Additionally, a series of Spearman correlations
with the inferential analysis without having to change were conducted between EM, IM, and (i) pre-
the theoretical structure of the questionnaire and sustainability scores, (ii) post-sustainability scores,
without removing/changing items. and (iii) ∆ scores, representing the difference between
pre and post-sustainability scores. This aimed to
assess the relationship between intrinsic and extrinsic
motivation and participants’ eco-friendly habits. A
significant negative correlation was identified
between pre-scores and the extrinsic motivation (EM)
construct (Figure 7; S = 83142, p-value = 0.047, 𝜌 = -
0.231).
Figure 6: Final model with significant standardized. Figure 7: Correlation between participants’
sustainability habits score before the campaign and their
Path STDYX Estimators S.E. score in the extrinsic motivation construct at the end of
IM ⇒ TA 0.876* 0.044 the campaign.
IM ⇒ BC 0.030 0.126
EM ⇒ TA 0.030 0.130 Importantly, no correlation was found between pre-
EM ⇒ BC 0.832* 0.054 scores and the overall score, suggesting that the
BC ⇔ TA 0.940* 0.221 campaign’s appreciation is independent of
BC ⇒ FB 0.386* 0.109 participants’ previous habits in sustainable
transportation. Finally, no correlations were observed
TA ⇒ FB 0.902* 0.074
between the other constructs and (i) pre-scores, (ii)
TA ⇒ OA 0.691* 0.157
post-scores, and (iii) ∆ scores. Then we confronted
FB ⇔ OA 0.491* 0.099
participants’ scores in the extrinsic motivation (EM)
Table 1: Standardized estimators between factors. and intrinsic motivation (IM) constructs (N = 117).
*𝑝 < 0.001 Also, in this case, we had to opt for a non-parametric
test since data did not distribute normally. A Wilcoxon
We used R v4.3.15 to run the analyses. Participants’ test highlighted a significant difference between the
overall score in the questionnaire indicates a good
5 https://cran.r-project.org/bin/windows/base/ 6 Supplementary material
189
two constructs, with IM being higher than EM (Figure
8; W = 539, p-value < 0.001).
To answer RQ2, we analyzed participants’
responses to the question "What did you like the most
about the campaign?" Forty-nine (49) participants did
not reply to the open question (or replied "all" or
"nothing"). Notably, a significant number of
participants emphasized the reward aspects (Reward
group, N = 27), while others expressed a preference for
the initiative’s underlying idea, such as promoting
sustainable mobility (Initiative group, N = 27). The
remaining 14 participants’ replies focused on other
elements, such as the competition, and traveling close Figure 9: Participants’ overall scores, divided per age
to nature. We sought to explore whether users’ and gender.
preferences had any impact on their scores in the
overall score and the intrinsic motivation (IM) and
extrinsic motivation (EM) constructs. First, a Mann- 6. Discussion
Whitney U test was employed to assess potential
differences in the overall score between the two The mobility campaigns involved a total of 362
groups (initiative, reward). The results indicated no participants, producing a collective 33,987 sustainable
significant difference in the overall score between the trips, for a total of 157,928 sustainable Kms, and
two groups (W = 293, p-value = 0.219). reducing CO2 emissions for a total of 34,8t. Analyzing
the questionnaires administered at the beginning and
end of the AIR-BREAK campaign, we found an increase
in terms of eco-sustainability by participants. In fact,
users reported significantly higher sustainability
scores at the end of the campaign. Furthermore, we
found that in the total number of users who completed
the post campaign questionnaire, the construct of
intrinsic motivation significantly outperforms that of
extrinsic motivation. This indicates that (RQ1) the AIR-
BREAK campaign is effective in promoting eco-
sustainable behaviors and particularly incentivizing
people to use more sustainable means in their daily
travel. Overall, the results obtained seem to indicate a
Figure 8: Participants’ scores in the IM and EM. balance of the two to produce both a behavioral change
and an appreciation of the application, resulting in a
Subsequently, a rank-transformed ANOVA was consequent long-term behavioral change expressed by
conducted (one factor between preference, 2 levels: the users and an overall appreciation of the campaign.
initiative, reward; one factor within subscale, 2 levels: As noted in the literature [20, 21], not necessarily the
IM, EM). The results revealed a significant main effect implementation of game design elements produces an
for the subscale factor (F1,104 = 9.792, p-value = 0.002), effective gameful system. In analyzing participants’
while no significant main effect was observed for the preferences toward the campaign elements, we
preference factor (F1,104 = 0.234, p-value = 0.630). identified two prevalent groups: those who showed
Tukey post-hoc test7 confirmed the difference particular interest in the rewards (virtual and
between EM and IM (IM M = 4.296, SD = 0.729; EM M = otherwise) related to the campaign, and those who
3.454, SD = 1.381; t = -3.129, p-value = 0.0023). liked the idea related to the initiative the most (RQ2).
Additionally, no significant interaction effect was Despite the division into the two groups, the data
present (F1,104 = 0.014, p-value = 0.907). Finally, to showed no differences in the total score, as well as
answer RQ3, we conducted a 2x4 rank-transformed indicating that in both cases IM is higher than EM. This
ANOVA to examine potential differences in overall suggests that the design of the gameful system does
scores based on participants’ age and gender (2 not make participants overly attached to rewards. Due
between factors: age, 4 levels: "<20", "20-35", "36-50", to the lack of a main effect of the group, as well as the
"51-70"; gender, 2 levels: female, male). The results absence of an interaction effect between the two
revealed no significant main effect for both age and factors analyzed, we can not identify specific elements
gender (age: F3,109 = 2.149, p-value = 0.0982; gender: in the campaign that motivated users more than others
F1,109 = 1.2246, p-value = 0.271), and no significant in adopting sustainable mobility behaviors. In
interaction effect between the two (F3,109 = 1.790, p- addition, we found no differences in campaign
value = 0.153) (Figure 9). appreciation (total score) by age and gender of
participants, indicating how the initiative is
appreciated by different demographic groups (RQ3).
As noted in the literature, it is possible that age and
gender moderate the appreciation for different aspects
of the gameful design [ 35, 34]. The absence of
7 Emmeans library
190
differences in appreciation of the initiative’s design also established the groundwork for enduring
and gamified system is an excellent finding, given that behavioral transformations. The emergence of local
the campaign aims to raise awareness of sustainability communities championing sustainable mobility and
with a broad demographic. We did find, however, how the highly positive feedback from participants suggest
EM correlates negatively with sustainability score at the project’s potential for a more extensive societal
the beginning of the campaign. This suggests that the influence. As urban centers persist in addressing the
presence of game elements may indeed capture the complexities of mobility, the AIR-BREAK project offers
attention of those who were not particularly inclined valuable lessons on the efficacy of gamified
to use environmentally sustainable means of interventions in fostering environmentally conscious
transportation. It also suggests that they do not lead to behaviors, opening avenues for future innovations in
phenomena such as the over justification effect [45, sustainable urban mobility initiatives.
46] in those users who were already showing high
levels of sustainability instead.
7.1. Lesson learned
6.1. Limitations The insights gained from the AIR-BREAK campaign
provide a foundation for future initiatives aiming to
It is necessary to point out that the reported results promote sustainable behaviors. Moving forward, the
have limitations. First of all, data in the literature positive correlation between intrinsic motivation and
report that for factor analyses, a size larger than 200 sustainability scores suggests a focus on designing
participants is recommended [47]. Furthermore, the gamified systems that tap into users’ inherent
sample size was found to be non-homogeneous motivations. Furthermore, understanding the diverse
between different analyses. Unfortunately, not all preferences of participants — some favoring rewards
users responded to either the pre- or post-campaign while others valuing the initiative itself — indicates
questionnaire, leading to varying sample sizes based the potential for tailored campaign elements to engage
on the analysis conducted. Regarding pre-post analysis a broader audience. To leverage these findings, future
in campaigns, it should be noted that few users campaigns could implement adaptive gamification
responded to the pre-campaign questionnaire, strategies, tailoring elements to individual preferences
reducing the reliability of this data. Although there is to maximize engagement. Additionally, considering
data on CO2 savings produced, behavioral change and the lack of significant differences in campaign
future behavior are inferred from items and not from appreciation based on age and gender, future
behavioral analysis, hence it is not possible to clearly initiatives might adopt inclusive design principles,
and linearly determine whether users were moving ensuring accessibility and appeal across diverse
sustainably even before the campaign. demographic groups. Addressing limitations, future
campaigns could strive for larger and more
homogeneous sample sizes, employing robust
7. Conclusions and future methodologies for pre-post analyses. The integration
works of behavioral analysis alongside self-reported data
could provide a more comprehensive understanding of
the sustained impact of gamified interventions on
In this paper, we introduced AIR-BREAK, a project
participants’ mobility choices.
designed to promote sustainable behaviors within the
general population, with a particular focus on
encouraging eco-sustainable mobility practices. AIR- Acknowledgements
BREAK has exhibited considerable success in
facilitating a positive shift towards more sustainable
This work is supported by the AIR-BREAK project
daily travel methods. The mobility campaigns
funded through the ERDF Urban Innovation Actions
developed and implemented in this project garnered
2020 UIA 05-177. We also acknowledge the support of
significant participation, resulting in a noteworthy
the PNRR ICSC National Research Centre for High
increase in sustainable trips, and a reduction in CO2
Performance Computing, Big Data and Quantum
emissions. The findings indicate that the gamified
Computing (CN00000013), under the NRRP MUR
approach, incorporating intrinsic motivation and tool
program funded by the NextGenerationEU.
attractiveness, has proven effective in motivating
individuals to embrace and maintain environmentally
friendly transportation choices. The questionnaire References
analyses revealed an overall improvement in
participants’ eco-sustainability scores. Notably, the [1] A. Vesco, F. Ferrero, Handbook of Research on
campaign’s effectiveness transcends demographic Social, Economic, and Environmental Sustainability
characteristics such as gender and age, indicating its
in the Development of Smart Cities, 2015.
broad appeal and impact. It’s worth mentioning that
[2] P. Cabanelas, G. Parkhurst, N. Thomopoulos, J. F.
while a significant number of participants identified
Lampón, A dynamic capability evaluation of
rewards as a positive aspect of the campaign, this did
not compromise their intrinsic motivation. In emerging business models for new mobility,
summary, the AIR-BREAK project has not only Research in Transportation Business Management
successfully met its initial goals of encouraging 47 (2023) 100964.
sustainable mobility throughout the campaign but has [3] R. Giesecke, T. Surakka, M. Hakonen,
Conceptualising mobility as a service, in: 2016
191
Eleventh International Conference on Ecological [15] A. Marconi, G. Schiavo, M. Zancanaro, G. Valetto,
Vehicles and Renewable Energies (EVER), 2016, pp. M. Pistore, Exploring the world through small green
1–11. doi:10.1109/EVER.2016.7476443. steps: Improving sustainable school transportation
[4] D. König, J. Eckhardt, A. Aapaoja, M. Sochor, J.and with a game-based learning interface, in:
Karlsson, Business and operator models for Proceedings of the 2018 International Conference
Mobility as a Service (MaaS). Deliverable 3 to the on Advanced Visual Interfaces, AVI ’18, Association
MAASiFiE project, Brussels -Belgium., 2016. for Computing Machinery, New York, NY, USA,
[5] B. Verplanken, S. Orbell, Attitudes, habits, and 2018.
behavior change, Annual review of psychology 73 [16] B. Cowley, J. Moutinho, C. Bateman, A. Oliveira,
(2022) 327–352. Learning principles and interaction design for
[6] S. A. Petersen, I. Petersen, P. Ahcin, Smiling earth- ‘green my place’: a massively multiplayer serious
raising awareness among citizens for behaviour game, Entertainment Computing 2 (2011) 10.
change to reduce carbon footprint, Energies 13 doi:10.1016/j.entcom.2011.01.001.
(2020) 5932. [17] B. Orland, N. Ram, D. Lang, K. Houser, N. Kling, M.
[7] M. Lindenau, S. Böhler-Baedeker, Citizen and Coccia, Saving energy in an office environment: A
stakeholder involvement: a precondition for serious game intervention, Energy and Buildings 74
sustainable urban mobility, Transportation (2014) 43–52. URL:
Research Procedia 4 (2014) 347–360. https://www.sciencedirect.com/
[8] J. Koivisto, J. Hamari, The rise of motivational science/article/pii/S0378778814000747.
information systems: A review of gamification doi:https://doi.org/10.1016/j.
research, International Journal of Information enbuild.2014.01.036.
Management 45 (2019) 191 – 210. [18] M. Shiraishi, Y. Washio, C. Takayama, V.
[9] S. Di Dio, E. Lissandrello, D. Schillaci, B. Caroleo, A. Lehdonvirta, H. Kimura, T. Nakajima, Using
Vesco, I. D’Hespeel, Muv: A game to encourage individual, social and economic persuasion
sustainable mobility habits, in: G. M., A. M., S. H. techniques to reduce co2 emissions in a family
(Eds.), Games and Learning Alliance - 7th setting, in: Proceedings of the 4th International
International Conference, GALA 2018, Proceedings, Conference on Persuasive Technology, Persuasive
Lecture Notes in Computer Science, Springer ’09, Association for Computing Machinery, New
Publishing Company, United States, 2019, pp. 60– York, NY, USA, 2009. URL:
70. doi:10.1007/978-3-030-11548-7_6, games and https://doi.org/10.1145/1541948.1541967.
Learning Alliance. GALA 2018 ; Conference date: doi:10.1145/1541948.1541967.
05-12-2018 Through 10-02-2019. [19] J. J. Lee, E. Matamoros, R. Kern, J. Marks, C. de Luna,
[10] J. Hamari, J. Koivisto, H. Sarsa, Does gamification W. Jordan-Cooley, Greenify Fostering sustainable
work? – a literature review of empirical studies on communities via gamification, in: CHI ’13 Extended
gamification, in: 2014 47th Hawaii International Abstracts on Human Factors in Computing Systems,
Conference on System Sciences, 2014, pp. 3025– CHI EA ’13, Association for Computing Machinery,
3034. doi:10.1109/ HICSS.2014.377. New York, NY, USA, 2013, p. 1497–1502.
[11] M. Ferron, E. Loria, A. Marconi, P. Massa, Play&go, [20] J. Koivisto, J. Hamari, The rise of motivational
an urban game promoting behaviour change for information systems: A review of gamification
sustainable mobility, Interaction Des. Archit. J 40 research, International Journal of Information
(2019) 24–25. Management 45 (2019) 191–210.
[12] S. Gabrielli, R. Maimone, P. Forbes, J. Masthoff, S. [21] K. Seaborn, D. I. Fels, Gamification in theory and
Wells, L. Primerano, G. Bo, M. Pompa, L. Haverinen, action: A survey, International Journal of human-
Designing motivational features for sustainable computer studies 74 (2015) 14–31.
urban mobility, in: CHI EA 2013 - Extended [22] F. Lieder, T. Griffiths, Helping people make better
Abstracts on Human Factors in Computing Systems, decisions using optimal gamification., in: CogSci,
volume 2013-April, Association for Computing 2016.
Machinery, 2013, pp. 1461–1466 [23] L. Wahlgren, P. Schantz, Exploring bikeability in a
doi:10.1145/2468356.2468617, 31st Annual CHI suburban metropolitan area using the active
Conference on Human Factors in Computing commuting route environment scale (acres),
Systems:, CHI EA 2013 ; Conference date: 27-04- International journal of environmental research
2013 Through 02-05-2013. and public health 11 (2014) 8276–8300.
[13] J. Bie, M. Bijlsma, G. Broll, H. Cao, A. Hjalmarsson, [24] M. Gallo, M. Marinelli, Sustainable mobility: A
F. Hodgson, P. Holleis, Y. Houten, K. Jacobs, J. review of possible actions and policies,
Koolwaaij, D. Kusumastuti, M. Luther, Move better Sustainability 12 (2020) 7499.
with tripzoom, International Journal on Advances in [25] P. Folco, L. Gauvin, M. Tizzoni, M. Szell, Data-driven
Life Sciences 4 (2012) 125–135. micromobility network planning for demand and
[14] R. Kazhamiakin, A. Marconi, M. Perillo, M. Pistore, safety, Environment and Planning B: Urban
G. Valetto, L. Piras, F. Avesani, N. Perri, Using Analytics and City Science 50 (2023) 2087–2102.
gamification to incentivize sustainable urban [26] R. Kazhamiakin, E. Loria, A. Marconi, M. Scanagatta,
mobility, in: 2015 IEEE First International Smart A gamification platform to analyze and influence
Cities Conference (ISC2), 2015, pp. 1–6. citizens’ daily transportation choices, IEEE
doi:10.1109/ISC2. 2015.7366196. Transactions on Intelligent Transportation Systems
192
22 (2021) 2153–2167. models: Tests of significance and descriptive
doi:10.1109/TITS.2021.3049792. goodness-of-fit measures, Methods of
[27] E. D. Mekler, F. Brühlmann, K. Opwis, A. N. Tuch, psychological research online 8 (2003) 23–74.
Disassembling gamification: the effects of points [42] G. Petri, C. G. von Wangenheim, A. F. Borgatto,
and meaning on user motivation and performance, Meega+: an evolution of a model for the evaluation
in: CHI’13 extended abstracts on human factors in of educational games, INCoD/GQS 3 (2016) 1–40.
computing systems, 2013, pp. 1137–1142. [43] J. Brooke, et al., Sus-a quick and dirty usability scale,
[28] R. N. Landers, K. N. Bauer, R. C. Callan, Gamification Usability evaluation in industry 189 (1996) 4–7.
of task performance with leaderboards: A goal [44] B. Muthén, D. Kaplan, A comparison of some
setting experiment, Computers in Human Behavior methodologies for the factor analysis of non-
71 (2017) 508–515. normal likert variables, British Journal of
[29] A. Anderson, D. Huttenlocher, J. Kleinberg, J. Mathematical and Statistical Psychology 38 (1985)
Leskovec, Steering user behavior with badges, in: 171–189.
Proceedings of the 22nd international conference [45] E. L. Deci, R. Koestner, R. M. Ryan, A meta-analytic
on World Wide Web, 2013, pp. 95–106. review of experiments examining the effects of
[30] R. Khoshkangini, G. Valetto, A. Marconi, M. Pistore, extrinsic rewards on intrinsic motivation.,
Automatic generation and recommendation of Psychological bulletin 125 (1999) 627.
personalized challenges for gamification, User [46] M. R. Lepper, D. Greene, R. E. Nisbett, Undermining
Modeling and User-Adapted Interaction 31 (2021) children’s intrinsic interest with extrinsic reward: A
1–34. doi:10.1007/ s11257-019-09255-2. test of the" overjustification" hypothesis., Journal
[31] A. Bucchiarone, S. Bassanelli, A. Marconi, How to of Personality and social Psychology 28 (1973) 129.
foster sustainable behaviors through multi- [47] T. A. Kyriazos, et al., Applied psychometrics: sample
campaigns rewarding mechanisms: The air-break size and sample power considerations in factor
experience, Sustainability 15 (2023) 5198. analysis (efa, cfa) and sem in general, Psychology 9
[32] A. Bucchiarone, S. Bassanelli, M. Luca, S. (2018) 2207.
Centellegher, P. Cipriano, L. Giovannini, B. Lepri, A.
Marconi, Play&go corporate: An end-to-end
solution for facilitating urban cyclability, IEEE
Transactions on Intelligent Transportation Systems
(2023).
[33] S. Bassanelli, N. Vasta, A. Bucchiarone, A. Marconi,
Gamification for behavior change: A scientometric
review, Acta Psychologica 228 (2022) 103657.
[34] J. Koivisto, J. Hamari, Demographic differences in
perceived benefits from gamification, Computers in
Human Behavior 35 (2014) 179–188.
[35] A. C. T. Klock, I. Gasparini, M. S. Pimenta, J. Hamari,
Tailored gamification: A review of literature,
International Journal of Human-Computer Studies
144 (2020) 102495.
[36] L. Zahedi, J. Batten, M. Ross, G. Potvin, S. Damas, P.
Clarke, D. Davis, Gamification in education: A
mixed-methods study of gender on computer
science students’ academic performance and
identity development, Journal of Computing in
Higher Education 33 (2021) 441–474.
[37] J. B. Ullman, P. M. Bentler, Structural equation
modeling, Handbook of Psychology, Second Edition
2 (2012).
[38] M. Carlsson, E. Hamrin, Evaluation of the life
satisfaction questionnaire (lsq) using structural
equation modelling (sem), Quality of Life Research
11 (2002) 415–426.
[39] L.-t. Hu, P. M. Bentler, Cutoff criteria for fit indexes
in covariance structure analysis: Conventional
criteria versus new alternatives, Structural
equation modeling: a multidisciplinary journal 6
(1999) 1–55.
[40] B. M. Byrne, Structural equation modeling with EQS
and EQS/Windows: Basic concepts, applications,
and programming, Sage, 1994.
[41] K. Schermelleh-Engel, H. Moosbrugger, H. Müller,
et al., Evaluating the fit of structural equation
193