Enhancing Sustainable Mobility Through Gamified Challenges: Evidence from a School-Based Intervention

Abstract

Promoting behavioral change in mobility is essential for sustainable urban development. This study evaluates the effectiveness of gamified challenges in fostering sustainable travel behaviors among high school students and teachers within the High School Challenge (HSC) 2024 campaign in Lecco, Italy. Over a 13-week period, participants tracked their commuting habits via gamified mobile application, Play&Go, that awarded points for sustainable mobility choices and introduced weekly challenges. Using behavioral (GPS-based tracking) and self-report data, we assessed the influence of challenge types, player characteristics (HEXAD Player Types, Big Five traits), and user experience evaluations on participation, retention, and behavior change. The results show that challenges, particularly those based on walking distances and framed as intra-team goals, significantly enhanced user engagement and contributed to improved mobility behaviors during participants’ free time. Compared to the 2023 edition without challenges, the 2024 campaign achieved better retention. HEXAD Player Types were more predictive of user appreciation than Personality Traits, though these effects were more evident in subjective evaluations than actual behavior. Overall, findings highlight the importance of tailoring gamified interventions to users’ motivational profiles and structuring challenges around SMART principles. This study contributes to the design of behaviorally informed, scalable solutions for sustainable mobility transitions.

1. Introduction

Human behavior is a primary driver of the environmental crises currently confronting our planet [1,2]. The rapid urbanization, widespread pollution, and deforestation are severely impacting the ecosystem and leading to a decline in biodiversity [3,4,5]. Addressing these challenges requires not only technological and structural solutions, but also widespread behavioral change, which has been identified as a critical component of climate mitigation strategies by the Intergovernmental Panel on Climate Change [6]. Transportation represents a key sector where behavioral changes can significantly impact environmental sustainability. As one of the largest contributors to greenhouse gas emissions, the transportation sector requires urgent action to encourage greener mobility behaviors [7]. In this context, urban policymakers face the double responsibility of guaranteeing for citizens the possibility to access mobility while simultaneously working to minimize the social, economic, and environmental impact of mobility systems [8,9]. Balancing between citizens’ well-being and sustainability needs calls for a comprehensive strategy that optimizes current mobility resources and integrates innovative services, ultimately trying to develop a sustainable and cohesive mobility ecosystem [10,11].

Adopting sustainable transportation options such as walking, cycling, public transit, and car sharing offers a wide range of benefits, from reducing pollution and greenhouse gas emissions to improving public health and urban livability [12,13]. Nevertheless, promoting a shift toward more eco-friendly travel modes remains an important challenge. Citizens’ habits, comfort preferences, and misconceptions about alternative mobility options often hinder this transition towards greener means of transportation [9,14]. Enhancing public awareness about the available mobility options, along with their environmental implications, is a critical step toward changing travel behaviors [15]. Individuals need to understand that their everyday mobility choices have collective consequences, impacting traffic congestion, emissions, and urban sustainability efforts [9,16]. To successfully promote greener mobility options, citizens must begin to see themselves as active contributors to a larger community effort aimed at achieving sustainable mobility goals.

In response to these challenges, cities are increasingly adopting digital and interactive technologies to engage citizens and encourage sustainable mobility behaviors. One promising approach is gamification, which applies game design elements to non-game contexts to motivate positive behavior change [17]. By making sustainable choices more engaging and rewarding, gamified systems can transform mobility habits and stimulate active participation in environmental initiatives [18,19]. Research across diverse fields, including transportation, health, environmental awareness, and tourism, has demonstrated the potential of gamified strategies to drive meaningful behavior change [20,21,22].

Recently, Bassanelli et al. [23] reviewed the literature to identify the initiatives that used gamification to promote sustainable behaviors in the population. The authors further analyzed the game elements adopted and how these represent persuasive strategies. Here, we describe some of the initiatives included in the literature review. Maca et al. [24], for example, studied the impact of Cyclers—a gamified smartphone application—to incentivize the use of bikes in the general population. They found that small financial incentives significantly enhanced the frequency of commuter cycling, especially when integrated with a gamification strategy. In contrast, gamification alone did not produce a statistically significant effect on cycling frequency. The authors used acknowledgments, points, competition, social pressure, and objectives (classification according to Toda et al.’s taxonomy [25] of game elements.) to promote the use of sustainable mobility among citizens over four weeks. On the other hand, other studies, such as those by Khoshkangini et al. [26] and Weber et al. [27], used gamified initiatives to promote sustainable mobility, obtaining positive results. Khoshkangini et al. [26] compared the use of recommended system-driven challenges (a game element falling into the broader category of objectives according to Toda et al.’s taxonomy [25].) vs. the use of manually assigned challenges in the context of gamification for sustainable mobility in the general population. The findings indicate that automatically generated challenges via a recommender system (RS) led to comparable or greater improvements in mobility-related behaviors than those manually assigned by game designers (non-RS). Moreover, rewards were found to be more efficiently utilized in the RS group compared to the non-RS group. In the second study, Weber et al. [27] concluded that the combination of gamification and smartphone-based logging adopted in Love to Ride significantly increased both participation and cycling activity. Participants using smartphone applications recorded a greater number of trips compared to those who did not, highlighting the effectiveness of digital tools in promoting active transportation. Social interaction and competitive elements further contributed to sustained engagement in cycling behavior.

Challenges are among the key game elements employed in gamified sustainable mobility campaigns. In contrast to other elements such as points and badges, which primarily function as feedback mechanisms, challenges offer users clear, structured, and personalized objectives [25]. According to motivational frameworks such as the SMART theory [28], well-defined goals can enhance individuals’ motivation and increase the likelihood of task completion. The SMART theory states that goals are more effective when they are: Specific (clearly defined and unambiguous), Measurable (quantifiable, allowing for assessment of progress and completion), Achievable (realistic and attainable), Relevant (aligned with overarching objectives), and Time-bound (associated with a defined deadline). In this regard, challenges represent an ideal mechanism for incorporating SMART goals into gamified systems. They allow the overall objective—encouraging more sustainable transport choices—to be broken down into smaller, actionable steps, thereby guiding users toward behavioral change through concrete and achievable tasks.

The use of challenges in gamified sustainable mobility campaigns is well established. Numerous studies have employed such game element, contributing to a growing body of knowledge on how challenges can influence travel behavior. For instance, in the work by Cellina et al. [29] and Kazhamiakin et al. [30], challenges were implemented to reduce car use and the associated CO₂ emissions. However, in both cases, the authors did not isolate the specific effects of the challenges nor clearly define the types of challenges applied. In contrast, Khoshkangini et al. [26] directly compared two challenge assignment methods—those generated via a recommender system and those manually assigned—and found that automatically assigned challenges were more effective in influencing behavior. Despite the increasing use of challenges in promoting sustainable mobility, there remains limited understanding of their specific effects on user behavior and motivation, as well as which types of challenges are most effective in achieving long-term modal shifts.

In this study, we examined how different types of challenges influenced user participation and retention in the HSC 2024 mobility campaign, and compared it to the 2023 initiative, which did not have challenges. HSC 2024 is a gamified initiative targeting high school students and their teachers in the Lecco area of Italy. Over 12 weeks, participants tracked their mobility behavior, contributing to their teams’ experience points and standings on a public leaderboard (a detailed description of the campaign is provided in Section 2). In this edition, we introduced a variety of challenge types to increase user engagement and reduce dropout rates. Our objective was to assess the effectiveness of these different challenge formats and investigate whether individual differences, specifically Personality Traits and Player Types, influenced their impact. Prior research in gamification suggests that Player Types can significantly affect users’ appreciation for and motivation to engage with various game elements [31,32], potentially making certain elements more or less suitable for a given population. Since the effect of individual game elements may also depend on the broader system design, we sought to explore how these elements interact with one another, with users’ intrinsic motivation, individual characteristics, and behavioral indicators. This broader analytical framework allowed us to interpret the effectiveness of the challenges in the context of the application as a whole.

This paper contributes to the literature on gamification and sustainable mobility by examining how specific game elements, particularly challenges, and individual differences, such as Personality Traits and Player Types, shape users’ motivation, engagement, and active participation in a real-world mobility campaign. Our research questions (RQs) therefore are the following:

RQ1: To what extent were the challenges introduced in the campaign effective in enhancing users’ participation and retainment in the mobility campaign?

RQ2: To what extent do users’ individual differences (i.e., Player Type and Personality Trait) influence their participation in the campaign and the challenges?

RQ3: To what extent are the users’ evaluations of game elements (i.e., experience points, leader boards, and challenges) related to their overall motivation and individual differences?

RQ4: To what extent are behavioral indicators related to users’ evaluation of and approach to the campaign?

In addition to assessing the impact of different challenge types on participation and retention, this study also seeks to contribute to a broader understanding of how individual differences influence engagement in gamified sustainability initiatives. We compare two widely used models: the Big Five [33] and the HEXAD scale [34]. While both frameworks can capture individual differences that influence engagement in gaming systems, they stem from different theoretical backgrounds. The Big Five offers a broad, domain-general model derived from personality psychology, while the HEXAD was developed specifically for gamification and focuses on users’ motivational preferences in relation to game elements. Prior research has shown that Personality Traits can predict users’ preferences for specific game elements [34,35] and support sustained motivation [36], while Player Types have been shown to enhance immediate user engagement [37], and some research find them to more directly capture users’ interactions with game features [34]. By integrating both models, we explore whether they offer complementary or overlapping insights and assess which more effectively predicts engagement in our context. This comparison can inform the design of more personalized and effective gamified interventions, (e.g., tailoring challenges, feedback, and team dynamics to user profiles), to better support sustainable mobility behavior.

The rest of the document is organized as follows. In Section 2, we describe the High School Challenge (HSC) initiative, including the challenge system and the general structure of the campaign. Section 3 explains recruitment methods, demographics, and student and teacher involvement. Section 4 explains the tools used for data collection and interaction, including the Play&Go app, surveys, and personality scales. Section 5 presents our findings, structured according to the research questions (RQs) outlined above. In Section 6, we interpret the results in light of the current literature and discuss theoretical and practical implications. This section also addresses limitations and directions for future research. Finally, in Section 7 we summarize the key findings and highlight the potential for more targeted and effective gamification strategies in the field of sustainable mobility.

2. High School Challenge Campaign

The High School Challenge (HSC) is an educational and behavioral intervention campaign designed to promote sustainable and active mobility among high school students. The competition targets entire school classes and leverages gamification elements to encourage daily low-impact commuting behaviors, including walking, cycling, public transport use, and carpooling. The first application of the campaign was in 2022 for the municipality of Lecco, and in 2023 for the municipality of Ferrara.

Class registration is facilitated by a designated reference teacher, who submits the team’s details via an institutional account. Teams must comprise students from a single class, with a minimum of 10 students or the entire class for smaller cohorts. Classes achieving at least 90% participation receive an initial bonus of 300 Eco-Leaves points. Teachers may also join their class team, though they can only be included in a single team. Participants engage through the Play&Go mobile application, a digital platform developed to track sustainable trips in real time. Each student’s sustainable trip earns Eco-Leaves points, which contribute to their class’ standing in the HSC rankings.

The competition spans 3 months, during which students track their trips by selecting the transportation mode and enabling GPS tracking through the Play&Go app. The app does not track trips automatically in the background; instead, participants must manually activate tracking at the start of each journey and specify the mode of transport used. This design choice, aimed at preserving user privacy, ensures that each recorded trip reflects an intentional act of engagement with the campaign. Once activated, each journey undergoes a two-step validation process. First, an automated check uses smartphone GPS data to calculate position and speed, verifying minimum trip length and compatibility with the declared transport mode. Second, a structural validation step cross-references public transport infrastructure and schedule data to confirm the trip’s plausibility. Although this system may result in some trips being missed if users forget to activate tracking, such manually recorded trips are considered ecologically valid indicators of active participation. This aligns with common practices in digital behavior change research, where system activity (e.g., logins, recorded events) is used to assess user participation and retention.

Points are awarded based on distance traveled, transport mode, and the number of active participants per team, with active mobility modes (walking and cycling) receiving the highest scoring coefficients per kilometer. Throughout the challenge, real-time weekly and global team leaderboards are available via the Play&Go App. These leaderboards display the current rankings of all participating teams based on the Eco-Leaves they have earned.

Beyond daily tracking, the competition incorporates weekly and thematic challenges designed to maintain engagement and stimulate peer collaboration. These challenges offer additional bonus points and are tailored to each team’s previous mobility behavior. The final standings are determined by the average Eco-Leaves points per team. Winning teams receive collective prizes, including educational experiences, recreational activities, and environmental excursions. Special awards are also conferred to teams for achievements in active mobility, consistency in participation, and carbon emission reduction. The HSC serves as a replicable model for municipalities and territories seeking to promote sustainable mobility through school engagement, real-time data collection, and community-based incentives. In this study, we compare the 2024 HSC campaign with its 2023 counterpart to assess the impact of the newly introduced challenges on participants’ behavior and retention. Both campaigns are identical in design, with the only difference being the addition of challenges in the 2024 edition. Below is a description of the game elements included in the campaigns.

2.1. Experience Points

The Mobility Score is calculated based on participants’ mobility behavior: users receive a specific amount of experience points (MS_m,d) based on the chosen means of transportation (m) and the distance traveled (d). It is an absolute measure, meaning it reflects only the individual’s actions, without being influenced by how many people are in their team, nor by point bonuses (i.e., initial team bonus, challenges, and survey bonuses). Participants do not see this score—it runs quietly in the background, simply tracking their mobility behavior. See

Table A1. Mobility Score points awarded for each km traveled with different means of transportation. To obtain the MS_m,d of a trip, the value in the table (m) is multiplied by the number of kilometers traveled (d). In order to obtain the Eco-Leaves awarded, the result needs to be divided by N—i.e., the number of players in the user’s group. * For carpooling, the number of MS points does not depend on the number of kilometers traveled, but rather on the number of passengers. Drivers are awarded 64 MS points for each passenger on each trip, provided there are a minimum of two people in the car. Trips must be at least one km long. Passengers are awarded 128 MS points for each trip. Only participants in the campaign can be considered passengers, since the app uses a QR code to validate the carpooling option.

Means of Transportation	MS Points per km
Foot	256
Bike and bike sharing	128
Bus	48
Train	24
Carpooling	64 *

in the Appendix A for more details on the conversion of users’ sustainable trips into the Mobility Score.

On the other hand, there are the Eco-Leaves (EC) points, which are calculated from the Mobility Score (MS_m,d) but adjusted according to the size of each participant’s group (N). The Eco-Leaves for each trip are calculated with the formula $EC=MS\mathrm{m},\mathrm{d}/N$. This makes Eco-Leaves a relative indicator. The idea is to level the playing field: regardless of a group’s size, each team should have the same opportunities to complete challenges and compete on the leaderboard. That is why Eco-Leaves are the score that participants see—they are used as feedback and motivation throughout the experience. To avoid unwanted behaviors, the MS points awarded—and therefore the EC collected—are progressively reduced.

According to this system, if everyone in a group walks the same distance, their group will earn the same number of Eco-Leaves as any other group that did the same—regardless of how many people are in each group. So, even though the Mobility Score offers a more direct measurement of each person’s sustainable behavior, the Eco-Leaves also reflect how difficult the challenges are in spite of the group size, making things fairer and potentially more engaging.

For example, Participant_01 walks 2 kilometers and earns 30 experience points in the Mobility Score. They are part of a group with six members, so their contribution is divided among the group, resulting in five Eco-Leaves (30 ÷ 6). Now consider Participant_02, who also walks 2 kilometers and earns the same 30 points. However, they are in a larger group of 15 people, so their contribution only results in two Eco-Leaves (30 ÷ 15). This system ensures that the total number of Eco-Leaves earned by each group remains balanced, regardless of the group’s size.

2.2. Challenges

Challenges in the HSC are structured tasks that require teams to achieve specific goals within a one-week period, demanding sustained engagement and effort. Upon successful completion, teams are awarded a fixed amount of Eco-Leaves.

These challenges can focus either on in-game performance metrics (e.g., earning a defined number of Eco-Leaves) or on behavioral outcomes related to sustainable mobility (e.g., tracking a minimum number of kilometers traveled on foot).

To effectively harness the motivational potential of gamification, challenge goals are personalized based on each team’s historical mobility patterns and performances. Engagement is optimized when challenge difficulty aligns with players’ abilities [38]: overly simple tasks risk boredom, while excessively difficult ones can lead to frustration. Consequently, challenges are carefully calibrated to require incremental improvements in mobility behavior while matching the team’s current performance level.

HSC incorporates both Single-Team and Couple of Teams challenges. Single-Team Challenges fall into two categories: (1) Absolute Increment Challenges that require a team to achieve a cumulative performance within a week (e.g., “Walk at least 180 km this week to earn a 250 Eco-Leaves bonus”), and (2) Repetitive Performance Challenges that require the repetition of a behavior across multiple days (e.g., “Walk at least 30 km per day for three days to earn a 250 Eco-Leaves bonus”).

Couple of Teams Challenges include both competitive and cooperative formats. A matchmaking algorithm pairs teams with similar performance levels by minimizing rank disparities. Challenge goals are then dynamically generated based on the past performances of the two matched teams. In cooperative challenges, both teams collaborate to meet a shared target. Individual contributions are not tracked and success is based solely on the joint outcome. If successful, both teams receive the same reward (e.g., “Team up with Team2 to walk at least 10 km together. Upon success, each team earns 250 Eco-Leaves”). In competitive time-based challenges, both teams aim to be the first to achieve a fixed target. The faster team wins the prize (e.g., “Race Team2 to walk 10 km. The first team to reach the goal wins 250 Eco-Leaves”).

These diverse challenge formats not only encourage healthy competition and collaboration but also foster long-term behavioral changes aligned with sustainable mobility goals.

2.3. Leaderboard

The leaderboard is based on each team’s total number of Eco-Leaves. Together with the challenges, it mixes cooperation and competition, a game modality often effective in motivating users [39,40]. The leaderboard is updated in real time to keep participants engaged and motivated throughout the entire campaign, not just at the end. It displays all participating teams, so that every group can track their progress and even try to improve. The team Eco-Leaves leaderboard is used at the end of the campaign to award class-level prizes to be distributed among the winning teams.

3. Participants

The HSC campaign targets high school students and their teachers, with students comprising the majority of the participant sample. In the 2024 edition, all participants were affiliated with high schools located in the Lecco area of Italy. Through the school, participants were informed about the initiative and could register a team with a minimum of 10 and a maximum of 30 participants. In the 2024 edition, 500 students and teachers registered for the campaign, for a total of 34 teams. In the 2023 edition, HSC Lecco had 44 teams and 694 registered users.

To answer RQ1, we analyzed data from all the users that actively participated in the campaign (average of active players per week, M = 210.58, SD = 81.06). On the other hand, to answer RQ2, RQ3, and RQ4, we relied on the participants who answered all the three questionnaires. Participation in the surveys was voluntary, and participants could complete any combination of the three administered questionnaires (i.e., the initial questionnaire, the individual differences questionnaire, and the final questionnaire). As illustrated in Figure A1, 362 participants completed the initial questionnaire, 89 completed the individual differences questionnaire, and 88 completed the final questionnaire. A total of 60 participants completed all three surveys. Of those 60 participants, we excluded the outliers concerning Player Types, Personality Traits, and the SuScores (an index of their mobility habits; see Section 4.1), for a final pool of 49 participants. Outliers were identified by excluding participants whose scores were higher or below two standard deviations from the mean. The demographics of the final sample are presented in

Table 1. Descriptive statistics by role (Teacher or Students).

	Teacher (N = 9)	Students (N = 40)	Overall (N = 49)
Age
Mean (SD)	53.3 (9.55)	15.1 (0.709)	22.1 (15.5)
Median [Min, Max]	57.0 [37.0, 63.0]	15.0 [14.0, 17.0]	15.0 [14.0, 63.0]
Gender
Female	7 (77.8%)	30 (75.0%)	37 (75.5%)
Male	2 (22.2%)	9 (22.5%)	11 (22.4%)
Prefer Not to Say	0 (0%)	1 (2.5%)	1 (2.0%)

. To assess the representativeness of the final sample, we compared it to the broader group of initial survey respondents (N = 362) across demographic and behavioral variables. Welch’s two-sample t-tests showed no significant differences in age, school commuting distance, or sustainable mobility scores. Pearson’s Chi-squared tests revealed no significant differences in gender distribution, and a marginally non-significant difference in school role distribution. These findings suggest that the final sample is broadly comparable to the initial sample in key characteristics, supporting its suitability for inferential analysis.

4. Materials and Methods

Throughout the campaign, users completed three questionnaires: one assessing users’ self-reported mobility habits at the beginning of the campaign, one investigating users’ Player Types and Personality Traits, and the last one on users’ self-reported mobility habits at the end of the campaign, containing also questions on participants’ appreciation and motivation for the campaign and the game elements (i.e., Eco-Leaves, challenges, and the leaderboard).

4.1. Initial Questionnaire on Sustainable Mobility

At the beginning of the campaign, participants were asked to complete a brief questionnaire covering their demographics (i.e., age and gender) and mobility habits. The mobility section included questions about their preferred means of transportation: (1) for the home-school trip and (2) during their free time. Additionally, participants reported the frequency of use of various modes of transportation—namely, car, carpooling, train, bus, bike, bike sharing, and walking (3) for the home–school trip and (4) in their free time (see Figure A2 and Figure A3 for bar graphs of self-reported data for each mean of transportation). These frequencies were assessed using a 5-point Likert-like scale (1 = less than once a month, 5 = more than four times a week). From responses to items (3) and (4), a Sustainability Score (SuScore) was computed using the following formula:

$$SuScore=\left[\right(6-car)+(sustainable\_means)/6]/2$$

“Sustainable_means” represents the total score in the items from all of the sustainable transportation. The formula reflects the goal of the campaign, which is to encourage the use of green transportation. By averaging the scores of eco-friendly options, while also including the inverse score of car usage, we aim to operationalize this goal into a single variable. The SuScore serves as an indicator of how sustainable a participant’s mobility habits are, with higher values reflecting more sustainable behavior. The initial questionnaire, as well as the SuScore, were already utilized in a previous study on a similar initiative [10].

4.2. Individual Differences: HEXAD Player Types and Big Five Personality Traits

To better understand participants’ appreciation of and engagement with the HSC campaign and its associated gamified app, we analyzed their Player Types alongside their Personality Traits. Incorporating both Player Types and Personality Traits in the customization of gamification represents a powerful strategy for enhancing user engagement.

Gamification Player Types (HEXAD Scale). Individual differences in motivation toward gamified experiences were assessed using the Gamification User Types HEXAD Scale [34], translated in Italian [41]. The HEXAD is a 24-item questionnaire that categorizes users into six types based on their dominant motivational drives in gameful systems: Philanthropists, who are motivated by purpose and helping others; Socializers, who seek social interaction and connection; Free Spirits, who value autonomy and creativity; Achievers, who are driven by mastery and goal accomplishment; Disruptors, who are oriented toward innovation and challenging the status quo; and Players, who are primarily motivated by extrinsic rewards. Participants responded to each item on a 7-point Likert scale ranging from 1 (strongly disagree) to 7 (strongly agree). For each user type, a composite score was computed by averaging the relevant items, following the scoring procedure outlined in the original validation study. The HEXAD scale has demonstrated good psychometric properties, including construct validity and internal consistency, across various domains. In our sample, internal consistency (Cronbach’s alpha) for each type was as follows: Philanthropists ($\alpha =0.80$), Socializers ($\alpha =0.79$), Free Spirits ($\alpha =0.77$), Achievers ($\alpha =0.78$), Disruptors ($\alpha =0.70$), and Players ($\alpha =0.85$).

Personality Traits (BFI-10). Broader personality dimensions were measured using the Italian validated version of the 10-item short version of the Big Five Inventory (BFI-10; [42,43]). This instrument provides brief yet valid assessments of the five major Personality Traits: Extroversion, Agreeableness, Conscientiousness, Emotional Stability (the inverse of Neuroticism), and Openness to Experience. Each trait is assessed using two items, with responses rated on a 5-point Likert scale (1 = strongly disagree, 5 = strongly agree). Scores were calculated by averaging the two items for each dimension. The BFI-10 is a widely used measure in large-scale survey research due to its brevity and adequate reliability for group-level analyses.

4.3. Final Questionnaire on Sustainable Mobility

At the end of the campaign, participants completed a final questionnaire that assessed their mobility habits post-intervention, along with a user experience (UX) section. The mobility habits section was similar to the initial questionnaire (see Section 4.1) and included questions on the frequency of use for the various means of transportation. From responses to the items in the mobility section, the SuScores post-campaign were calculated. The UX section included (1) a question on participants’ overall motivation; (2) a question on the pleasantness of each game element (i.e., Eco-Leaves, challenges, and the leaderboard); (3) a question on participants’ motivation related to each game element; (4) a question on the perceived impact of the campaign on home–school trips (referred as Behavioral school in the figures in the Results Section) and in the free time (referred as Behavioral free time in the figures in the Results Section). In the results, we will refer to (2) as game elements’ appreciation, as it refers to how much users deemed the game element pleasant during the campaign; we will refer to (3) as game elements’ motivation as the item referred to how much users perceived the game element motivated them throughout the campaign. Similarly to the initial questionnaire, the final questionnaire was already utilized in a previous publication on a similar initiative [10].

4.4. Eco-Leaves and Mobility Score

In the analysis of the 2024 HSC initiative, we decided to include both the Eco-Leaves and the Mobility Score. Based on our hypothesis, the Mobility Score should better estimate participants’ mobility habits, as it represents an absolute indicator of one’s sustainable behavior. We verified this assumption by correlating participants’ Mobility Score with their Eco-Leaves and saved CO₂ (see

Table 2. Correlations among Mobility Scores, Eco-Leaves, and the CO₂ saved ( *** $p<0.001$).

	Eco-Leaves	Mobility Score
Eco-Leaves
Mobility Score	0.75 ***
CO2	0.55 ***	0.72 ***

). The CO₂ savings for each trip were calculated based on the difference between the CO₂ emissions of a car and those of the transport mode actually used. This provides a measure of the trip’s environmental sustainability. On the other hand, users see Eco-Leaves as feedback for their performance. Feedback plays a crucial role in gamification, as it helps support adopting and maintaining new behaviors [44,45,46]. Considering the importance of the two Experience points, we included them both in our analyses and tested the difference between them.

5. Results

5.1. Results Overview

The following section presents the results of the High School Challenge campaign, structured around the four research questions. Our goal was to evaluate the impact of gamified challenges on participation and behavior, examine the role of individual differences in user engagement, and explore the relationships between users’ experiences, motivation, and self-reported behavior change.

All analyses were quantitative. The entire statistical workflow was conducted in R version 4.1.3, with the exception of the survival analysis, which was conducted in Python using the lifelines package. We used Welch’s t-tests and Chi-squared tests to assess sample representativeness and demographic comparability. Pearson correlations were used to examine associations between individual differences and subjective evaluations of the game elements. Mixed-effects linear models were employed to analyze repeated measures related to user engagement in challenges. Finally, survival analysis was used to model retention and dropout dynamics over time in relation to psychological traits.

We begin by addressing overall campaign efficacy, using both objective (GPS-tracked trips) and self-reported indicators of sustainable mobility to examine changes in behavior, with a focus on differences between home–school and free-time travel. Differences over time were tested using descriptive comparisons across survey waves and sustainability score trends.

To answer RQ1, we analyze how different types of challenges (e.g., Single Team vs. Couple of Teams; distance-based vs. points-based) influenced participation and retention. These effects were assessed through mixed-effects linear models (implemented via the lme4 package in R), which accounted for within-participant repeated measures. We further explored participant retention dynamics by comparing current dropout patterns with those from the previous year’s campaign using survival analysis in Python.

For RQ2 and RQ3, we examine the role of individual differences—namely, HEXAD Player Types and Big Five Personality Traits—in shaping user engagement and responses to specific game elements, including challenges, Eco-Leaves, and leaderboards. These associations were investigated through Pearson correlations, allowing us to identify patterns between motivational/psychological traits and subjective user evaluations and appreciation scores.

Finally, to address RQ4, we assess how behavioral indicators are related to participants’ motivational profiles and overall evaluations of the campaign. We use correlational analyses to explore these associations and identify whether higher appreciation or motivation is linked to more sustainable behavior during the campaign.

5.2. Campaign Efficacy

In the 2024 HSC campaign, 500 participants, including students and teachers, enrolled to take part in the High School Challenge initiative. Data from the Play&Go mobile application showed that, during the 13 weeks, they tracked over 25,000 sustainable journeys, covering more than 143,000 sustainable kilometers—of which 4.8% were by bike and another 13.4% walking, which are zero-impact means of transportation (see Figure 1 and Figure 2). The data are aligned with the 2023 results, where participants tracked 34,170 trips (53.9 per person in 2023 vs. 50.49 in 2024) for a total of 143,315 sustainable kilometers. For a graphic representation of the data, see Figure A4 and Figure A5 in Appendix A.

Sustainability score. To evaluate the efficacy of the campaign in promoting the use of sustainable means of transportation, we analyzed the difference between the SuScores at the beginning and at the end of the campaign. The analysis was run in R v 1.4.3.2 using the t.test function for paired samples, as the data distributed normally. The results showed no significant difference in the scores for the home–school trips (pre SuScore $M=2.748, SD=0.517$; post SuScore $M=2.747, SD=0.567$; $t=0.025, df=48, p=0.980$) and a significant difference in the SuScores related to participants’ free time ($t=-3.010, df=48, p=0.004$) with the scores being higher at the end of the campaign (pre SuScore $M=2.202, SD=0.454$; post SuScore $M=2.388, SD=0.616$). The observed increase of $0.186$ points in the average SuScore corresponds to a small-to-moderate effect size (Cohen’s d = $0.34$). This indicates a modest yet significant improvement in participants’ mobility habits during their free time. Since the SuScore ranges from 1 to 5 and reflects increased use of sustainable transportation methods and decreased car usage, this change suggests a positive, albeit limited, shift toward more sustainable mobility behaviors. Figure A2 and Figure A3 in the Appendix A provide a graphical overview of these results.

Motivation. Participants found the campaign, on average, motivating (General motivation $M=3.531, SD=1.043$). For what concerns the game elements, users found Challenges to be the most motivating one ($M=3.510, SD=1.157$), followed by the Leaderboard ($M=3.306, SD=1.294$) and Eco-Leaves ($M=3.265, SD=1.151$). All the game elements received a positive evaluation, as they scored higher than the neutral value of 3.

Appreciation. In terms of appreciation, participants deemed Challenges the most pleasant game element ($M=4.020, SD=0.924$), followed by Eco-Leaves ($M=3.816$, $SD=1.014$) and the Leaderboard ($M=3.755, SD=1.011$). All the game elements received a positive evaluation, as they scored higher than the neutral value of 3.

5.3. RQ1: Effectiveness of Challenges in Enhancing Participation and Retention in Sustainable Mobility Campaign

To assess how challenges can enhance people’s participation and retention in the sustainable mobility campaign, we first visualized weekly trends in players’ Mobility Scores. As shown in Figure 3, the average player score increased following the introduction of challenges in Week 4. While temporary drops in engagement were observed around Weeks 5 and 9, corresponding to Italian school holidays, the overall trend suggests a positive effect of the challenges on sustainable choices. See Appendix A Figure A6 for the plot representing the average Eco-Leaves, which follows a similar trend.

We then statistically examine the influence of types of challenges and challenges’ type of goals on users’ participation using as dependent varibales both types of Experience Points: Mobility Score, the absolute indicator of individual sustainable behavior, and Eco-Leaves, the more relative score adjusted for group size and used for gamified feedback. As shown in Table 2, the Mobility Score represents the best indicator of users’ sustainable behavior, as it shows a stronger correlation with the CO₂ (calculated based on the difference between the CO₂ emissions of a car and those of the transport mode used) saved than Eco-Leaves. Nonetheless, the two metrics are not perfectly correlated with each other, which reinforces the decision to analyze both: the Mobility Score provides a purer behavioral measure, while Eco-Leaves capture user-facing motivational effects shaped by the campaign’s social mechanics.

To statistically examine these effects, we fitted linear mixed-effects models with random intercepts for participant (playerId) to control for interindividual variability. All models were estimated using the lme4 package [47], with p-values obtained via lmerTest [48], and post-hoc comparisons conducted using estimated marginal means from emmeans [49].

The first model tested whether participation differed depending on whether users participated in Single Teams or in Couple of Teams challenges (either cooperatively or competitively). The effect of Challenge Type was significant, for both Mobility Score (${\chi }^2\left(2\right)=15.08$, $p<0.001$) and Eco-Leaves (${\chi }^2\left(2\right)=12.87$, $p=0.002$). Post hoc comparisons (Tukey-adjusted) revealed that there was no difference in participation (both Eco-Leaves and Mobility Score) between Cooperative and Competitive Couple of Teams challenges (lower $p=0.505$). Single Team challenges lead to higher Mobility Score and Eco-Leaves compared to Competitive Couple of Teams challenges (higher $p=0.007$). However, while Single Team challenges lead to higher Eco-Leaves compared to Cooperative Couple of Teams challenges ($p=0.028$), the difference was only marginal in Mobility Scores ($p=0.054$).

We then tested whether different Types of Goals (i.e., Eco-Leaves, Distance in Km) influenced engagement. The effect was significant for both Mobility Score (${\chi }^2\left(1\right)=48.31$, $p<0.001$) and Eco-Leaves (${\chi }^2\left(1\right)=24.60$, $p<0.001$), revealing that participation was higher when players were given a challenge goal in terms of distance (in Km) to be covered walking in a week compared to a target number of Eco-Leaves to reach in the same span of time.

Exploratory Comparison with Prior Campaign (2023 vs. 2024)

To explore whether the introduction of challenges may have influenced overall player retention throughout the campaign, we descriptively compared the weekly count of active players (i.e., registered users who recorded at least one sustainable trip during a given week of the campaign) in the 2024 campaign to that of a previous campaign conducted in the same schools in 2023, when no challenges were introduced. These schools had participated in an earlier edition of the campaign in 2022. This meant that the participants in 2023 and 2024 were not entirely new to the initiative.

As illustrated in Figure 4, the number of active participants declined over time in both years, as expected in this type of intervention. However, the rate of decline appeared less steep in 2024 compared to 2023.

Although there were fewer active players at the start of the 2024 campaign, the two groups converged in their final weeks, ending with nearly identical numbers of active participants by Week 13. These comparisons are exploratory and should be interpreted with caution. However, the pattern shown in Figure 4, supported by the Eco-Leaves data, suggests that the presence of weekly challenges in 2024 may have sustained engagement over time and buffered the typical decline observed in such interventions.

Despite a lower number of active players at the beginning of the 2024 campaign, the two years converged in their final weeks, resulting in a nearly identical number of active participants by Week 13. While these comparisons are exploratory in nature, the pattern shown in Figure 4, supported by the Eco-Leaves data, suggests that the presence of weekly challenges in 2024 may have helped sustain engagement over time, potentially buffering the typical decline observed in such interventions. (https://www.smartcommunitylab.it/lecco-playgo-high-school-mobility-challenge/. Last accessed: 9 July 2025).

However, we acknowledge that factors beyond the challenges themselves may have influenced retention outcomes. Specifically, it is possible that participants in 2024 were more motivated due to their prior exposure to the campaign in previous years. Additionally, it is possible that a higher proportion of individuals who had already engaged with the campaign in previous editions chose to participate in 2024, making them inherently more motivated or familiar with the process. Together with the introduction of new challenges, these factors may have contributed to the observed retention patterns.

To complement these visual trends with more robust statistical comparisons, we conducted a survival analysis comparing the 2023 and 2024 cohorts. The analysis was performed using JupyterLab (version 4.0.11), modeling retention as a time-to-event process in which the event of interest is user disengagement. Each participant was considered at risk of disengagement each week, and the survival probability reflects the likelihood of continued participation over time. A Kaplan–Meier estimator [50] was used to estimate weekly retention probabilities, treating each active week as a survival event. As shown in Figure 5, participants in the 2024 campaign showed higher retention rates than those in 2023. The difference between the survival curves was found to be statistically significant according to the log-rank test ($p<0.001$), indicating that the 2024 campaign produced longer-lasting engagement.

To better understand this effect and account for differences in user behavior, we applied a Cox proportional hazards model [51] that uses average weekly mobility score and Eco-Leaves as standardized covariates, includes campaign year as a binary indicator, and tests for interaction effects. The Cox model estimates dropout risk as a function of specific user-level characteristics and allows us to assess both behavioral and contextual predictors of retention. The results show that a higher average weekly mobility score was significantly associated with better retention. Specifically, a one standard deviation increase in mobility score meant a 16% reduction in dropout risk ($HR=0.84$, $p<0.001$). It is important to note that, although the mobility score was significantly associated with retention, participants were not exposed to this metric directly during the campaign. Therefore, this association likely reflects underlying behavioral patterns or engagement levels rather than the mobility score directly influencing retention. The number of weekly Eco-Leaves was not a significant predictor ($HR=1.02$, $p=0.53$). As in the previous model, participation in the 2024 campaign was still significantly associated with a lower risk of quitting compared to 2023, corresponding to a 10% lower risk ($HR=0.90$, $p<0.001$), even after accounting for behavioral factors.

To examine whether the effect of these behavioral predictors differed across campaign years, interaction terms between year and each behavioral variable were included. However, neither the interaction between campaign year and weekly mobility ($HR=1.04$, $p=0.35$) nor the interaction between campaign year and weekly Eco-Leaves ($HR=0.99$, $p=0.84$) reached statistical significance. These results suggest that, while individual mobility behavior plays an important role in predicting retention, the impact of this behavior on dropout risk does not differ significantly between the two campaign years.

The model demonstrated good predictive accuracy, with a concordance index of $0.82$, indicating a strong ability to discriminate between users who would remain active and those who would drop out.

5.4. RQ2: Association Between Users’ Individual Differences (i.e., Player Type and Personality Trait) and Participation with the Campaign and the Challenges

To better understand the relevance of individual differences in the context of gamified behavior change interventions, we examined both Personality Traits (Big Five) and gamification-specific motivational orientations (HEXAD Player Types). While these two frameworks share some conceptual overlap, they capture distinct dimensions of individual variation, one rooted in general personality structure, the other in users’ motivational profiles within interactive systems. As such, examining both is essential for evaluating which framework offers greater predictive utility in the context of gamified sustainable mobility campaigns.

The descriptive statistics and boxplot representing the distributions of each of the Big Five Personality Traits and HEXAD Player Types are reported in the Appendix A (

Table A2. Descriptive statistics for Big Five Personality Traits and HEXAD Player Types.

Trait/Type	Mean (SD)
Big Five Personality Traits
Extraversion	3.10 (0.93)
Agreeableness	3.27 (0.91)
Conscientiousness	3.63 (0.84)
Emotional Stability	2.49 (0.92)
Openness	3.36 (0.86)
HEXAD Player Types
Philanthropist	5.65 (0.90)
Socialiser	5.16 (1.01)
Free Spirit	5.67 (0.82)
Achiever	5.72 (0.84)
Disruptor	3.61 (1.19)
Player	4.72 (1.11)

, Figure A7 and Figure A8). Among the Big Five traits, Conscientiousness ($M=3.63$, $SD=0.84$) and Openness ($M=3.36$, $SD=0.86$) displayed the highest mean scores, whereas Emotional Stability ($M=2.49$, $SD=0.92$) was the lowest. This suggests that participants tend to see themselves as organized and open to new experiences but report lower levels of emotional resilience.

As for the HEXAD types, Achiever ($M=5.72$, $SD=0.84$), Free Spirit ($M=5.67$, $SD=0.82$), and Philanthropist ($M=5.65$, $SD=0.90$) were the most endorsed, indicating a motivational orientation centered around mastery, autonomy, and purpose. In contrast, Disruptor ($M=3.61$, $SD=1.19$) was the least endorsed type. These patterns align with prior research on adolescents, where intrinsic and prosocial motivations tend to be more salient. Notably, Player ($M=4.72$, $SD=1.11$) scored lower than most intrinsic types, reflecting a more moderate attraction to extrinsic rewards.

As shown in Figure 6, several significant associations emerged between user types and Personality Traits. Philanthropist correlated positively with Extraversion, Conscientiousness, and Agreeableness, reinforcing the prosocial and conscientious characteristics associated with this user type. Socializer exhibited a strong positive association with Agreeableness, Conscientiousness, and Extroversion which is coherent with its interpersonal orientation. Both Free Spirit and Achiever were significantly and positively correlated with Openness and Conscientiousness. Free Spirit was slightly more strongly associated with Openness, aligning with theoretical expectations, given their shared emphasis on creativity and autonomy. On the other hand, Achiever presented stronger correlations with Conscientiousness, consistent with the idea that achievement-oriented individuals are also more organized and goal-driven. High levels of Achiever trait were also associated with higher Emotional Stability.

Finally, both the Player and the Disruptor type were significantly correlated with Openness, a pattern that, while it may seem less intuitive, could reflect a shared attraction to novelty and stimulation. While Players seek engaging and varied reward structures, Disruptors may resonate with Openness through their desire for innovation and non-conformity.

To explore how individual differences such as Player Types and Personality Traits relate to sustainable mobility behaviors in our campaign, we computed Pearson correlation coefficients between the six HEXAD user types, the Big Five Personality Traits, and our Experience Points. These last scores included Eco-Leaves, Mobility Score, and their respective variants for Single Team, Competitive, and Cooperative Couple of Teams challenge types. The correlation matrix was computed using the rcorr() function from the Hmisc package [52] in R.

Table 3. Pearson correlations between HEXAD user types and mobility variables (split by scoring type; significant results in bold, * $p<0.05$).

(a) Eco-Leaves (EL)
	Overall	Single	Couple-Comp	Couple-Coop
Philanthropist	−0.01	−0.05	−0.06	−0.12
Socializer	0.06	−0.03	0.10	−0.04
Free Spirit	0.00	−0.06	−0.17	−0.14
Achiever	0.34 *	0.32 *	0.19	0.13
Disruptor	0.05	0.13	−0.07	0.03
Player	0.18	0.21	0.07	−0.01
(b) Mobility Score (MS)
	Overall	Single	Couple-Comp	Couple-Coop
Philanthropist	0.00	−0.03	0.04	−0.08
Socializer	0.05	−0.03	0.17	−0.04
Free Spirit	−0.01	−0.07	−0.21	−0.13
Achiever	0.21	0.20	0.04	−0.06
Disruptor	0.03	0.14	−0.12	0.00
Player	0.17	0.23	0.03	0.01

presents the correlations between HEXAD types and mobility outcomes. Among all relationships tested, the trait Achiever showed the only significant associations: it correlated with Eco-Leaves in general and with Eco-Leaves in the Single Team challenges; however, the correlations for this player type and the other challenge types were not significant. No other significant correlations emerged across the HEXAD types and sustainable mobility behaviors. Finally, no significant associations were found between Big Five Personality Traits and sustainable mobility behaviors (see Appendix A 

Table A3. Spearman correlations between Big Five Personality Traits and Mobility Variables (split by scoring type).

(a) Eco-Leaves (EL)
	Overall	Single	Couple-Comp	Couple-Coop
Extroversion	−0.03	−0.08	0.04	−0.11
Agreeableness	0.02	−0.05	0.12	0.06
Conscientiousness	0.23	0.19	0.20	0.10
Emotional Stability	0.06	0.05	0.04	0.06
Openness	0.12	0.12	0.04	0.00
(b) Mobility Score (MS)
	Overall	Single	Couple-Comp	Couple-Coop
Extroversion	−0.10	−0.12	0.05	−0.15
Agreeableness	−0.04	−0.13	0.12	0.01
Conscientiousness	0.13	0.06	0.15	−0.01
Emotional Stability	−0.06	−0.03	−0.08	−0.11
Openness	0.17	0.16	0.08	0.14

).

5.5. RQ3: Game Elements’ Evaluation and Individual Differences

To investigate the relationship between users’ evaluation of game elements and individual differences, we performed a Pearson correlation between the six HEXAD user types, the Big Five Personality Traits, and users’ answers to the UX section of the final questionnaire (i.e., overall motivation, game elements-related motivation, game elements appreciation).

Among the Big Five Personality Traits, only Emotional Stability showed significant results; participants who reported higher emotional stability also reported lower appreciation for Eco-Leaves and lower motivation related to challenges. See Figure 7B. Regarding the HEXAD Player Types, the overall motivation was positively correlated with both Achievers and Players. Notably, Players exhibited the highest association with appreciation for game elements, showing significant positive correlations with appreciation for Eco-Leaves (consistently with [34]) and challenges. However, these associations were limited to appreciation, as no significant correlations were observed with game element-related motivation (see Figure 7).

5.6. RQ4 Behavioral Indicators, Game Elements’ Evaluation, and Individual Differences

To assess the relationship between users’ behavioral indicators (i.e., delta sustainability score and perceived behavioral change—both during the home-school trips and in the free time) and users’ evaluation and approach to the campaign, we performed a Pearson correlation between the behavioral indicators, and users’ answers to the UX section of the final questionnaire.

In general, both the appreciation and motivation for game elements were positively and significantly associated with users’ self-reported perceived behavioral change—both during home–school trips and in their free time (see Figure 8). The highest correlations were observed for motivation related to Eco-Leaves, with home–school behavior and free time behavior. This was followed by motivation related to challenges (for home–school behavior, for free time behavior) and appreciation of Eco-Leaves (for home–school behavior, for free time behavior). An exception emerged with the appreciation of challenges, which did not show a significant correlation with home–school behavior, but did correlate significantly with free time behavior.

Users’ delta sustainability score during free time showed a significant positive relationship with all three motivational evaluations of game elements: Eco-Leaves, challenges, and leaderboard, supporting the alignment between users’ perceptions and actual behavior change (i.e., Behavioral free time). In contrast, the delta sustainability score for home–school trips did not show any significant correlations.

Furthermore, we analyzed the relationship between behavioral indicators and users’ individual differences (i.e., Player Types and Personality Traits) using Pearson correlation. A significant positive correlation was observed between participants’ perceived behavioral change (i.e., Behavioral free time) and the Player Type from the HEXAD framework ($r=0.33$; see Figure A9 in Appendix A for the full matrix of correlations). No significant relationships were found between the Big Five Personality Traits and behavioral indicators (see Figure A10 in Appendix A for the full matrix of correlations).

6. Discussion

This study examined the effectiveness of a gamified intervention employing challenges, the High School Challenge, in promoting sustainable mobility behaviors among high school students and teachers. We explored four key research questions: (1) the extent to which the introduction of gamified challenges influenced user participation and retention; (2) the role of individual differences (Player Types and Personality Traits) in modulating these effects; (3) the relationships between users’ evaluation of gamification elements and their motivation; and (4) the association between behavioral indicators and users’ perceptions of the campaign. Finally, we explanatorily investigated whether the Big Five Personality Trait or the HEXAD scale would be a better tool to determine how individual user differences influence engagement in gamified sustainability initiatives. The results offer promising evidence for the potential of gamified strategies to foster the transition toward sustainable mobility. Participants’ sustainable mobility behaviors during free time improved significantly over the campaign, and engagement was maintained more effectively than in a previous edition that did not include challenges. Challenges based on distance goals (e.g., walking kilometers) were more effective in promoting participation than those based on abstract point thresholds. Participants showed a preference for Single Team challenges over Cooperative and Competitive Couple of Teams challenges. In terms of user characteristics, Player Types showed stronger associations with appreciation and perceived impact than Personality Traits, with Achievers and Players particularly responsive to Eco-Leaves and challenges, although predominantly at the level of appreciation rather than motivation.

6.1. Theoretical and Practical Implications

The comparative analysis of the 2023 and 2024 editions of the campaign suggests an association between introducing challenges and maintaining user participation. Although a decline in participation is common over time in long-duration digital interventions [46], the challenges may have helped sustain participants’ interest. This finding aligns with temporal self-regulation theory, which shows that periodic reinforcement of goals helps sustain behavior over time by helping against loss of motivation [53]. In gamification research, challenges are regarded as goal-oriented prompts that help maintain attention, promote flow experiences, and reduce attrition in extended interventions [54]. Our results further support the idea that challenges are particularly effective when aligned with SMART goal principles [28]; that is, when they are specific, measurable, achievable, relevant, and time-bound. Walking-based challenges supported this structure and were associated with higher levels of participation than point-based goals, reinforcing recent findings that goal clarity and perceived attainability are key drivers of participation in mobility applications [55].

Moreover, we observed that Single Team challenges were more effective than Cooperative or Competitive Couple of Teams challenges in sustaining participation. One possible explanation is that Single Team formats reduce social comparison and possibly performance anxiety, factors known to weaken intrinsic motivation, especially in adolescents [56]. Additionally, intra-team challenges may simplify coordination and clarify responsibilities, making participation less demanding. While prior literature has shown the potential of competition and social incentives [57], our findings suggest that, in adolescence and in sustainability contexts, intra-group cooperation without external pressure may offer a more supportive environment for behavior change. This extends the literature by showing the importance of challenge structure in gamified interventions, particularly when aiming to sustain retention of sustainable mobility programs.

These findings also contributed to the understanding of how gamified systems can support user acceptance and participation in the context of sustainable mobility transitions. The consistent positive correlations between participants’ overall motivation and their evaluations of all core game elements, challenges, leaderboards, and Eco-Leaves showed that users perceived the campaign as engaging. This supports recent research indicating the importance of positive user experiences in encouraging the adoption of soft mobility modes such as walking, cycling, and public transport [58]. Designing interventions that align with users’ expectations appears central to facilitating behavioral change in this domain. This highlights the practical value of evaluating users’ motivational profiles and preferences toward specific game elements, enabling more personalized and engaging designs that can better sustain participation.

In addition to user experience, our findings offer insight into the role of individual differences in shaping gamified engagement. HEXAD Player Types, especially Achievers and Players, were more consistently associated with appreciation for game elements than Big Five Personality Traits, suggesting that understanding users’ motivation could be more useful for improving the design [34]. However, these correlations were limited to the appreciation rather than the motivation or behavior, indicating the need to distinguish between liking a gamified element and being driven by it [59]. This insight adds to ongoing discussions in the field and highlights the importance of considering different sides of users when designing technologies that aim to change behavior.

Finally, the results highlight the importance of contextual flexibility in supporting sustained behavioral changes. Behavioral changes were observed in participants’ free-time mobility, while no significant changes were found for home–school mobility. This may reflect not only the fact that free-time activities are less constrained by schedules or infrastructure, but also that many students were already commuting to school using sustainable modes such as walking or public transport before the campaign, leaving limited room for further improvement (see Figure A3). For those who did rely on cars, structural barriers such as limited access to public transit, long distances, or rigid timetables may have prevented behavior change even during the intervention. To support this interpretation, we include in the Appendix A.3 a breakdown of self-reported transport mode usage collected at baseline (see Figure A2 and Figure A3). These findings are consistent with Fogg’s Behavior Model [60], which shows that behavior change is most likely to happen when high motivation is matched with ability and well-timed triggers. Sustainable mobility interventions should therefore target domains where users can realistically act on their intentions, maximizing both engagement and behavioral impact. Furthermore, the observed behavioral patterns support the potential use of digital indicators (e.g., trip logs, activity frequency, challenge participation) as predictive tools for identifying user engagement levels and tailoring interventions accordingly. This suggests a promising direction for research focused on linking real-time behavioral data with user evaluations and campaign outcomes to optimize design decisions.

From a practical perspective, this study offers guidance for the design of gamified systems aimed at promoting sustainable mobility. First, gamified challenges should be based on clear and achievable goals, better if related to physical metrics such as distance traveled rather than abstract points. Second, designers should consider proposing Single Team-based competition formats, which had higher participation than cooperative or inter-team competition models in this study. Users may be more motivated by intragroup dynamics and clear team identity than by competition or externally imposed collaboration. Third, feedback mechanisms such as Eco-Leaves, which helps to visualize performance, are important for reinforcing sustainable behaviors. Fourth, gamification frameworks should prioritize player type profiles over broad Personality Traits when tailoring design features. Although we should be cautious because their ability to predict behavior is limited, Player Types are still useful for gamified activities since they are based on motivation, especially for school-age children. These findings offer practical tools for designing engaging, scalable interventions that foster behavioral shifts toward sustainable transport modes.

Beyond design recommendations, the findings and implementation experience from this study also inform strategies for scaling up gamified mobility interventions. We offer the following considerations to support policymakers, educators, and practitioners aiming to implement such initiatives on a broader level. Integrate with existing structures: Embedding gamified campaigns into school curricula, municipal sustainability plans, or regional mobility initiatives enhances legitimacy, visibility, and institutional support. Ensure digital accessibility: Reduce technological exclusion by providing equitable access to devices and optimizing apps to minimize battery and performance constraints—an issue we encountered in our own campaign, especially among users with older smartphones. Foster stakeholder collaboration: Align goals and resources through collaboration between schools, transport authorities, municipalities, and community organizations to sustain long-term engagement and ensure shared ownership. Tailor interventions to user motivation: Use motivational profiling (e.g., Player Types) to personalize content and adapt challenge difficulty over time. This approach proved especially important among adolescents. Focus on high-autonomy domains: Our results show that behavioral change was more likely in free-time mobility contexts—where participants have more control—than in constrained contexts like home–school commuting. Prioritizing domains where people can realistically act on their intentions may lead to greater impact.

6.2. Limitations and Future Research

While the findings are encouraging, it is important to acknowledge this study’s limitations, which can serve as guidance for future research.

One limitation concerns the use of self-reported behavior, which can be influenced by social desirability and recall bias [61]. Although we mitigated this issue by incorporating more objective measures, such as the Mobility Score calculated from GPS-based tracking via the Play&Go app, some outcomes, such as the Sustainability Score, were from questionnaires. Future research should work on improving behavioral measurement by integrating more digital logs and sensor-based data.

Moreover, although our study employed a 13-week pre–post design, which offers a solid foundation for observing changes in behavior over time, it did not include follow-up assessments beyond the campaign period. This limits our ability to determine whether the behavioral shifts observed during the intervention were temporary or enduring. Additionally, the study did not include a traditional randomized control group, which restricts the strength of causal claims. To partially address this, we compared participation and dropout dynamics with those observed in the 2023 edition of the campaign, which did not include gamified challenges. This comparison provided a meaningful benchmark to estimate the added value of the 2024 gamified design. Nonetheless, future research should incorporate both post-campaign follow-up assessments and controlled designs to better evaluate the long-term impact and isolate the specific contributions of individual intervention components.

It is important to note that the campaign was first introduced in 2022. Both the 2023 and 2024 implementations took place in the same schools that had already been exposed to the campaign. This initial deployment likely established a shared background of awareness and familiarity within the school communities, influencing participant engagement in subsequent years. While the individuals who participated in 2023 and 2024 may have differed in motivation or attitude, introducing a possible self-selection bias, such effects were likely present in both cohorts. Therefore, we argue that the conditions created by the 2022 launch contributed to a consistent context across years, supporting the comparability of the two campaigns. Nevertheless, future research should examine participant characteristics more closely and adopt study designs capable of disentangling these factors.

The generalizability of our findings is also influenced by the context of the study, as participants were all from Italian high schools. Cultural norms, infrastructural constraints, and demographic variables may all influence how gamified interventions are received. Although our setting has high ecological validity, replication across different populations and settings will be necessary to prove the robustness of these findings.

Another consideration involves the types of behaviors targeted. While in the intervention we assessed both more constrained (home–school commuting) and more flexible (free-time) mobility contexts, behavior change was seen mainly in the latter. This asymmetry indicates the importance of aligning intervention with types of mobility that give users enough autonomy and opportunity for change.

Moreover, our design included multiple game elements all playing a simultaneous role in the campaign, which makes it difficult to disentangle their individual effects. It remains unclear whether the observed outcomes were more driven by challenges, feedback mechanisms, social competition, or their interaction. While our multi-element approach reflects more the complexity of real-world interventions, future studies should consider isolating specific mechanics to better understand their contributions to user participation and behavior.

Finally, our sample may have been influenced by self-selection bias. Because only participants who completed both pre- and post-questionnaires were included, it is possible that individuals already likely to act in sustainable ways were overrepresented. Although this problem is not unique to our study, it shows that different strategies might be needed to reach and keep participants who start with low motivation or have different motivational types, like Disruptors.

Despite these limitations, this study provides a robust and ecologically grounded contribution to the literature on gamified sustainability interventions. The integration of behavioral, motivational, and individual-level data in a real-world campaign setting advances both theoretical understanding and practical design of tools aimed at fostering transitions toward more sustainable mobility practices.

7. Conclusions

This study provides evidence that gamified challenges, when well-structured and contextually appropriate, can effectively support the transition toward sustainable mobility. By using SMART goal-based tasks in a real-world school campaign, we observed increased participation, higher retention, and improvements in sustainable behaviors, particularly during free time, where users had more autonomy to act. Key design features such as Single Team formats and clear feedback mechanisms contributed to a positive user experience. Player Types were more useful than Personality Traits in predicting users’ appreciation of gamification, reinforcing the value of motivational profiling in intervention design. Ultimately, while this study is context-specific, it contributes to the broader objective of understanding how to design systems that facilitate the transition toward soft, shared, and sustainable modes of transportation, aligning users’ experiences with ecological and policy goals.