Social media engagement and pedestrian bridge utilisation: investigating impacts of anticipated regret and community enforcement visibility Open Access

Rapid urbanisation and increasing vehicular traffic have intensified concerns about pedestrian safety, especially in developing cities where road crashes remain a leading cause of injuries and fatalities (World Health Organization [WHO], 2018). In recognition of this global challenge, the Sustainable Development Goal (SDG) 3.6 explicitly targets a 50% reduction in road traffic deaths and injuries by 2030 (United Nations, 2015). Achieving this target requires not only infrastructural investments but also behavioural change strategies that encourage safer road-use practices.

Pedestrian bridges are among the most common structural interventions designed to separate pedestrian movement from vehicular traffic and reduce crash risk. However, evidence consistently shows that utilisation rates remain low despite significant investments (Sadeek et al., 2025). For instance, in Ghana, 93% of pedestrian fatalities along the highway occurred while attempting to cross roads at grade rather than using available footbridges (Damsere-Derry and Bofah, 2023). National statistics similarly indicate that over 73% of pedestrian crash casualties result from unsafe crossing behaviours, including the neglect of footbridges (Alimo and Agyeman, 2022). Comparable findings from other developing contexts show that many pedestrians still prefer risky at-grade crossings due to convenience, time pressure, or limited awareness of safety risks (Ackaah and Adonteng, 2011; Khatoon et al., 2013; Mukherjee, 2025).

This persistent gap between infrastructure provision and behavioural compliance underscores the importance of understanding the psychosocial and communicative drivers of pedestrian behaviour. Without addressing these behavioural factors, progress toward SDG 3.6 may be hindered, as infrastructure alone cannot guarantee safer mobility. Recent developments in digital communication provide new opportunities to address this challenge. Social media engagement (SME) in this study refers to individuals' active interaction with road-safety content on digital platforms, including activities such as viewing, reacting to, sharing or commenting on messages, which reflects both their cognitive processing and behavioural participation in online safety communication. Beyond general communication functions, SME has gained increasing relevance in transport and mobility research. Studies show that social media platforms serve as critical channels for disseminating traffic updates, safety alerts, hazard warnings and behaviour-change messages (Oviedo-Trespalacios et al., 2019; Alshareef and Ahn, 2022). Such platforms shape perceptions of road-use norms, influence compliance intentions and support community-level safety discourse.

Although SME has been widely examined in marketing, public health and risk-communication research, its role within transport and mobility studies is still emerging. Recent findings suggest that digital platforms can shape mobility attitudes, increase compliance with safety regulations and influence perceptions of road-use risk (Zhao et al., 2021). In pedestrian-safety contexts, social-media campaigns influence awareness, processing of safety cues and behaviour change, particularly in urban areas with high mobile-internet usage (Nicolls et al., 2025a, b; Faus et al., 2024). Despite the growing evidence of SME's influence on road-safety attitudes, little research has examined whether such engagement translates into concrete mobility behaviours such as pedestrian-bridge utilisation.

In urban-mobility contexts, SME has been shown to affect pedestrian visibility campaigns, distracted-walking behaviour and adherence to safety guidelines, demonstrating its behavioural influence beyond awareness alone (Faus et al., 2024; Oviedo-Trespalacios et al., 2019). However, it appears no empirical study has explored whether SME predicts the actual use of pedestrian infrastructure, leaving a major gap in transport-safety literature.

While these platforms can shape attitudes and awareness, it remains unclear whether engagement with road-safety campaigns translates into sustained behavioural change, such as consistent use of pedestrian bridges. This gap is especially salient in low- and middle-income countries, where digital adoption is high but evidence of behavioural impact remains limited.

A key cognitive–emotional mechanism (Huang et al., 2022) that may mediate this relationship is anticipated regret (AR), defined by the extent to which individuals expect to feel negative emotions when choosing an unsafe option. AR has been shown to motivate preventive behaviours in health and risk-related contexts (Lorimer et al., 2025), yet its influence on pedestrian safety decisions, particularly bridge utilisation, remains underexplored.

Beyond psychological mechanisms, the broader social environment also plays a crucial role in shaping pedestrian behaviour, particularly through the visibility of community enforcement measures such as wardens, fines, signage and official communication. These cues can reinforce behavioural norms by signalling accountability and heightening the perceived consequences of unsafe actions (Acemoglu and Wolitzky, 2020; Imam, 2022; Marx and Archer, 1971). Visible enforcement has shown measurable effects on compliance, yet its interaction with digital engagement remains theoretically underexplored.

Nevertheless, empirical research has yet to determine whether such visible enforcement enhances or diminishes the influence of SME and AR on pedestrians' crossing decisions. Although extensive research has addressed infrastructure design, traffic conditions and enforcement interventions in developing contexts (Chen, 2010; Kombonaah et al., 2025; Osei-Kyei and Chan, 2016; Sharma and Dehalwar, 2025), studies on digital road-safety campaigns still prioritise awareness creation rather than behavioural outcomes (Faus et al., 2024; Mishra and Dev, 2025; Nicolls et al., 2025a, b). Furthmore, almost no evidence connects SME directly to real-world pedestrian behaviour such as safe-crossing practices or bridge usage, revealing a substantial gap in transport-safety research.

Although AR strongly predicts preventive behaviour in other domains (Lorimer et al., 2025), its application to pedestrian decisions remains limited. Similarly, community enforcement visibility (CEV) despite its documented importance in reinforcing compliance through social norms and accountability (Acemoglu and Wolitzky, 2020; Imam, 2022) remains underexplored alongside digital communication and psychological mechanisms in pedestrian-safety settings. Therefore, this study addresses these gaps by examining how SME influences pedestrian bridge utilisation (PBU), with AR as a mediator and CEV as a moderator. The specific objectives are to investigate the:

1. Influence of SME on PBU and AR regarding unsafe crossing behaviour.

2. Impact of AR on PBU.

3. Mediating role of AR in the relationship between SME and PBU.

4. Moderating effect of CEV on the relationships among SME, AR, and PBU.

This study is grounded in Protection Motivation Theory (PMT), originally developed by Rogers (1975) to explain how individuals adopt protective behaviours in response to perceived threats. PMT has been widely applied in health communication, risk prevention and safety behaviour studies because it integrates both cognitive and emotional processes in decision-making (Doane et al., 2016; Floyd et al., 2000; Raj Sreenath et al., 2025). According to PMT, protective behaviour results from two main appraisal processes: threat appraisal and coping appraisal.

Threat appraisal involves evaluating the severity of a potential risk and one's vulnerability to it. In the context of pedestrian behaviour, road traffic accidents represent a highly severe and prevalent threat, particularly in urban settings with high vehicular traffic. Engagement with social media safety campaigns can heighten awareness of these risks and make the dangers of at-grade crossings more salient. Within this process, AR plays a crucial role as an emotional mechanism, amplifying the perceived severity of not using pedestrian bridges by evoking the expectation of negative emotions if harm occurs as a result of unsafe crossing (Lorimer et al., 2024; Richard et al., 1996).

Coping appraisal, on the other hand, concerns evaluating the effectiveness of the recommended protective action, one's ability to perform it, and the associated costs or barriers. PBU can be viewed as the recommended protective action, and its adoption depends on whether individuals perceive it as effective and feasible. Here, CEV acts as a contextual factor that strengthens coping appraisal. Visible enforcement through signage, fines or safety wardens signals the effectiveness of bridge use as a protective behaviour and reinforces perceptions of social accountability.

By integrating these components, PMT provides a useful framework for understanding how SME influences PBU through AR (threat appraisal) and how CEV conditions this relationship (coping appraisal). The theory suggests that individuals are more likely to adopt safe road-crossing behaviours when they both recognise the threat of accidents and believe that protective behaviours, such as using pedestrian bridges, are effective and socially reinforced.

Thus, PMT not only justifies the study's focus on psychological and contextual variables but also guides the hypothesised pathways: SME as a cue that raises threat appraisal, AR as a mediator that translates awareness into motivation, and enforcement visibility as a moderator that strengthens coping appraisal and subsequent behaviour change.

Social media has emerged as a central channel for disseminating safety information, accident reports and peer experiences, which collectively influence pedestrian decision-making (Nicolls et al., 2025a, b). Drawing on PMT, exposure to such information enhances both threat appraisal (perceptions of accident severity and vulnerability) and coping appraisal (belief in the efficacy of protective actions such as bridge use). When pedestrians engage with safety-related content, they become more aware of risks associated with unsafe crossings and more convinced of the utility of using pedestrian bridges (Rijal and Yilmaz, 2024; Schwebel et al., 2012). Thus, SME directly motivates protective behaviours, including bridge utilization. Based on the preceding discussion, the following hypothesis is proposed:

In addition to shaping cognitive risk appraisals, SME may also elicit affective responses such as AR. Social media platforms frequently circulate accident-related narratives, crash footage and near-miss incidents involving pedestrians (Chang et al., 2022a, b; Lee et al., 2022a, b), which have been shown to heighten emotional responses and influence safety-related intentions (Glik, 2007; Liu and Fan, 2020). Exposure to such content can prompt individuals to mentally simulate future scenarios in which failing to use a pedestrian bridge may lead to harm, thereby intensifying AR (Zeelenberg, 1999; Richard et al., 1996). Consistent with PMT, this reflects the coping-appraisal component, where individuals evaluate both the emotional and practical consequences of their behavioural choices (Brewer et al., 2016a, b). Consequently, SME may indirectly promote safer pedestrian behaviours by amplifying feelings of AR.

AR has been widely recognized as a strong predictor of protective behaviours across domains such as safe driving, vaccination, and risk avoidance (Sandberg and Conner, 2008; Xiong et al., 2024). In the pedestrian context, the expectation of regret from not using a bridge, particularly if it results in harm, may serve as an emotional motivator for adopting the safer option. Thus, AR could function as a key psychological mechanism that translates social media influence into behavioural change (Chitraranjan and Botenne, 2024). Drawing from the above arguments, the study proposes the following hypothesis.

Integrating these arguments, AR may be positioned as a mediator between SME and bridge utilisation. While social media provides information and raises awareness (Faus et al., 2024; Moharam et al., 2024; Nicolls et al., 2025a, b), it is often the AR that may drive pedestrians to act on this information. PMT supports this mechanism by recognizing the dual role of cognitive and affective processes in motivating protective behaviours (Ogbanufe and Pavur, 2022; Rogers, 1983). Therefore, it is proposed that SME not only directly influences pedestrians' bridge use but also indirectly affects it through the mediating role of AR. Accordingly, the study advances the following hypothesis.

While SME and AR could influence PBU, the extent of their impact may depend on the social and environmental context. In this regard, CEV such as the presence of local wardens, neighbourhood watchdog groups or public signage indicating enforcement plays a crucial role in shaping behavioural outcomes. This visibility also strengthens collective efficacy (Alehegn et al., 2025), which in turn promotes more effective informal social control and safer community behaviours (Nader et al., 2025). Drawing on PMT, enforcement visibility can be understood as an external coping resource that enhances individuals' perceptions that protective behaviours (e.g. using pedestrian bridges) are both expected and reinforced within the community.

Although SME increases awareness of risks and encourages safer crossing behaviours (Tyagi et al., 2025), these intentions may not always translate into actual behaviour due to competing pressures such as convenience, time constraints or habit. However, when community enforcement is visible, the perceived social cost of non-compliance rises (Sandt et al., 2016), making pedestrians more likely to act consistently with their intentions. Thus, enforcement visibility may strengthen the relationship between SME and bridge utilization by creating an external accountability mechanism that transforms awareness into action. Based on the preceding discussion, the following hypothesis is proposed.

Similarly, AR may motivate pedestrians to avoid risky behaviours by imagining the negative consequences of not using bridges. However, while internal emotional drivers such as AR may motivate protective intentions (Brewer et al., 2016a, b; Chitraranjan and Botenne, 2024), they are not always sufficient to overcome situational barriers such as convenience, accessibility, or environmental constraints (Sheeran et al., 2013). The presence of visible enforcement reinforces these emotional motivations by adding a tangible layer of deterrence, making the potential regret more salient and immediate. Under high enforcement visibility, pedestrians are more likely to act on AR and adopt safe practices. Drawing from the above arguments, the study proposes the following hypotheses.

Figure 1 and Table 1 illustrate the conceptual framework and provide a detailed description of the study constructs based on the theoretical foundations and hypothesis development.

Figure 1 presents the proposed relationships among SME, AR, CEV and pedestrians' bridge utilisation, highlighting the mediating and moderating mechanisms guiding the study model. More so, Table 1 complements the framework by defining each construct, along with its empirical justification.

The study was conducted in major urban centres in Ghana, where pedestrian activity and road-use patterns are heavily shaped by the physical and cultural environment. These cities are characterised by dense commercial zones, high-traffic arterial roads and mixed land-use feature pedestrian bridges located near markets, transport terminals, schools and industrial corridors. Such locations typically experience heavy congestion, frequent jaywalking and limited enforcement, which influence how pedestrians perceive the convenience and safety of bridge use. Culturally, informal crossing norms remain widespread, as many commuters prioritise speed and familiarity over formal road-safety infrastructure. Additionally, Ghana has a highly active social media landscape driven by strong mobile internet penetration, making digital platforms central to community communication, public sensitisation and road safety discourse. This geographical and cultural context is therefore essential for understanding how SME interacts with AR and CEV to shape PBU.

This study employed a quantitative, cross-sectional survey design to examine the relationships between SME, AR, CEV and PBU. A cross-sectional approach was appropriate for capturing behavioural patterns and perceptions from a large number of respondents at a single point in time (Kesmodel, 2018). Structural Equation Modelling with Partial Least Squares (PLS-SEM) was used because it is well-suited for predictive studies involving complex models with mediation and moderation, and it performs robustly with non-normal data (Sarstedt and Cheah, 2019). The target population comprised pedestrians in urban areas with pedestrian bridges, as they regularly decide whether to use bridges and are likely exposed to road safety campaigns on social media. A multi-stage sampling strategy was applied: first, urban areas with pedestrian bridges were purposively selected, followed by random selection of pedestrians to ensure demographic diversity. Following Wolf et al. (2013) guideline of at least 10 cases per structural path, the minimum sample size was 150. To enhance statistical power and address non-responses, 915 respondents were surveyed.

Data were collected using a structured questionnaire divided into two sections. The first captured demographics such as age, gender and education, along with bridge use and SME. The second measured the study constructs using five-point Likert scales (1 = strongly disagree, 5 = strongly agree), with items adapted from validated scales (Dessart et al., 2016; Dolan et al., 2019; Hasan et al., 2020; Jiang et al., 2023; Räsänen et al., 2007; Sozer and Merlo, 2013). A pilot test with 30 respondents ensured clarity and reliability. SME was measured with five items assessing interaction with safety content; AR with three items on perceived risks of unsafe crossing; PBU with five items on frequency and consistency of bridge use; and CEV with five items on perceived presence of wardens, signage, patrols, and penalties (see Table 2).

Each construct was modelled reflectively, and reliability and validity were evaluated through Cronbach's alpha, composite reliability (CR) and average variance extracted (AVE). Discriminant validity was further assessed using the Fornell–Larcker criterion and the heterotrait–monotrait (HTMT) ratio.

Data analysis was conducted using SmartPLS 4.1, following a two-step approach. First, the measurement model was assessed for internal consistency, convergent validity, and discriminant validity. Second, the structural model was evaluated to test the hypothesized relationships. Bootstrapping with 5,000 resamples was used to assess the significance of path coefficients. Additionally, the coefficient of determination (R²), effect sizes (f²) and predictive relevance (Q²) were reported. Mediation analysis was conducted to assess the role of anticipated risk, while moderation analysis tested the interaction effects of CEV. All results were presented with effect sizes and confidence intervals to meet rigorous academic standards.

Table 3 summarises the demographic characteristics and behavioural patterns of the 915 respondents. The largest group was aged 26–35 years (36.6%), followed by 18–25 years (31.7%), 36–45 years (21.6%), and above 45 years (10.1%). Gender distribution was fairly balanced, with 54% male and 46% female. Regarding education, 34% held a bachelor's degree, 32% had a diploma, 10% had postgraduate qualifications, and 24% had completed secondary education, indicating that most participants had attained some form of tertiary education. Patterns of PBU varied; 40% reported rarely using pedestrian bridges, 25% used them sometimes, 20% often and 15% always. These results suggest that despite awareness, consistent use of pedestrian bridges remains relatively low. In terms of SME with road safety campaigns, most respondents (60%) indicated they always engaged, 18.8% engaged often, while 13.1% engaged sometimes and 8.1% rarely. This demonstrates high levels of interaction with digital safety content, which may influence awareness and behavioural intentions.

Since the data for this study were collected using self-reported questionnaires, the potential for common method bias (CMB) was assessed. Following the recommendation of Podsakoff et al. (2003), Harman's single-factor test was conducted using exploratory factor analysis. The results indicated that no single factor accounted for the majority of variance, with the first factor explaining 45% (less than the 50% threshold), suggesting that CMB was not a significant concern in this study. Additionally, to further confirm the absence of bias, a full collinearity test was performed as recommended by Kock (2015). All variance inflation factor (VIF) values were below the cut-off value of 3.3, providing further evidence that common method variance was unlikely to distort the results.

To establish the reliability and validity of the study instruments, both construct reliability and discriminant validity were rigorously assessed (Chau, 1999; Hancock and Mueller, 2001). As presented in Table 4, the Cronbach's alpha values for all constructs ranged from 0.848 to 0.964, indicating strong to excellent internal consistency across the measurement items. In addition, Table 5 reports the CR and the AVE for each construct. All factor loadings exceeded 0.60, CR values were above 0.70, and AVE values surpassed the 0.50 threshold, thereby confirming satisfactory convergent validity (Henseler et al., 2016).

Discriminant validity was examined using both the Fornell–Larcker criterion and the HTMT ratio, two widely applied approaches in structural equation modelling (Henseler et al., 2015). According to Fornell and Larcker (1981), discriminant validity is achieved when the square root of each construct's AVE is greater than its correlations with other constructs. The results in Table 5 confirmed this condition, supporting the presence of discriminant validity. Moreover, HTMT values (reported in Table 5) did not exceed the recommended threshold of 0.85 (Henseler et al., 2016), with the highest value recorded at 0.84. Collectively, these results demonstrate that the measurement model possesses both adequate reliability and robust discriminant validity.

The model demonstrated a satisfactory fit, with SRMR values of 0.072 (saturated model) and 0.075 (estimated model), both below the recommended threshold of 0.08 (Henseler et al., 2016). In addition, the d_G, d_ULS and Chi-square values for the saturated model were lower than those for the estimated model, further supporting acceptable model fit (Henseler, 2017) as presented in Table 6.

The analysis of endogenous constructs shows that AR and CEV demonstrate strong predictive relevance and explanatory power. Stone-Geisser's Q² values for AR (0.760) and CEV (0.447) confirm predictive relevance, while R² indicates moderate explanation for AR (53.7%) and substantial explanation for CEV (76%). Predictive accuracy was satisfactory, with RMSE and MAE values showing consistency between predicted and observed outcomes (see Table 7). Effect size analysis further reveals that SME exerts the greatest influence on both PBU (f² = 1.036) and AR (f² = 1.158), far exceeding the 0.35 threshold for large effects. This establishes SME as the dominant predictor, directly promoting bridge use while indirectly influencing behaviour through AR. In comparison, CEV's effect on AR (f² = 0.125) was small-to-moderate, and AR's effect on PBU (f² = 0.033) was small. Overall, SME drives bridge use, with AR and CEV offering complementary support.

Table 8 and Figure 2 presents the results of the structural model assessment. The findings demonstrate strong support for all hypothesized relationships. First, SME significantly influences AR with a standardized path coefficient of β = 0.733, t = 14.715, p < 0.001, confirming H1. This suggests that increased engagement with social media safety campaigns heightens pedestrians' feelings of AR when they consider unsafe road behaviours. Similarly, H2 was supported, showing that SME has a significant positive influence on pedestrian bridge utilization (β = 0.938, t = 24.712, p < 0.001). This indicates that frequent exposure to and interaction with social media safety messages directly encourages pedestrians to adopt safer behaviours by using pedestrian bridges. The relationship between AR and PBU was also significant (β = 0.138, t = 2.797, p = 0.005), thereby supporting H3. Although the effect size is smaller compared to direct SME impact, this result implies that pedestrians' AR contributes positively to their likelihood of using pedestrian bridges.

In Table 8, the moderating role of CEV was also confirmed. For H5a, the interaction between SME and CEV significantly influenced PBU (β = 0.294, t = 4.978, p < 0.001), suggesting that community visible enforcement strengthens the positive effect of SME on bridge usage. Likewise, H5b was supported (β = 0.094, t = 2.006, p = 0.045), indicating that CEV also enhances the relationship between AR and PBU. Taken together, these results highlight that both cognitive (SME) and affective (AR) mechanisms significantly drive pedestrian bridge utilization, and these effects are further amplified by institutional and community-level enforcement visibility. This reinforces the importance of integrating digital campaigns with physical enforcement to achieve more sustainable behavioural change in pedestrian safety.

The total effect of SME on PBU was 1.04, slightly exceeding the standardized −1 to +1 range but acceptable in PLS-SEM, where total effects combine both direct and indirect pathways (Hair et al., 2017a, b; Sarstedt et al., 2022). This indicates that SME has an exceptionally strong overall influence on bridge use. However, the variance accounted for (VAF) was 10.8%, showing that only a small portion of this effect is explained by AR. As VAF below 20% suggests no mediation, the results confirm that AR does not meaningfully mediate the SME–PBU relationship (see Table 9).

This study aims to investigate the influence of SME on PBU, with AR serving as a mediating mechanism and CEV as a moderating factor. Before interpreting the findings, it is important to contextualise the study area. The research was conducted in major urban centres in Ghana, characterised by dense commercial activity, high pedestrian volumes, frequent jaywalking and limited enforcement of pedestrian-safety regulations. These features shape daily mobility patterns and strongly influence pedestrian decision-making, making the context especially relevant for examining behavioural responses to digital road-safety communication.

Firstly, the study reveals that SME significantly predicts pedestrians' AR. This means that when pedestrians actively engage with road safety content on social media platforms, they become more likely to anticipate negative emotions if they choose unsafe behaviours, such as crossing roads without using bridges. Supporting this, prior studies (Berger and Milkman, 2012; Lee and Hong, 2016; Phua et al., 2017) have demonstrated that social media campaigns can elicit affective responses that significantly shape behavioral intentions. Likewise, Richard et al. (1996) and Zeelenberg and Pieters (2007) highlight that AR is a strong motivator in discouraging risky behaviours and encouraging protective ones. This finding is consistent with evidence from health-behaviour and driver-safety studies, where engagement with online campaigns has been shown to evoke emotional reflection and enhance risk-avoidance across diverse contexts.

Second, the study shows that SME has a strong and direct positive influence on PBU. In other words, frequent exposure to and interaction with safety campaigns online significantly encourage pedestrians to adopt safer practices by choosing bridges. This finding resonates with prior studies that demonstrate the effectiveness of digital platforms in shaping health and safety behaviours (Mahapatra and Mishra, 2017; Alalwan et al., 2017). Similarly, Oviedo-Trespalacios et al. (2019) found that exposure to road safety messages through digital channels increases compliance with pedestrian safety measures, reinforcing the crucial role of online communication in behaviour change. However, compared to studies conducted in high-income settings with stronger enforcement cultures, the result suggest that SME may play an especially prominent role in Ghanaian and other developing-country urban areas where institutional enforcement is inconsistent. This highlights a context-specific pathway through which digital engagement compensates for weaker on-ground regulatory structures.

Thirdly, the study reveals that AR positively predicts PBU, though with a smaller effect compared to SME. This suggests that pedestrians who reflect on the potential negative consequences of unsafe crossing are more likely to adopt protective behaviours. This finding is consistent with research showing that AR drives people to avoid risky decisions and instead choose preventive options (Sandberg and Conner, 2008; Brewer et al., 2016a, b). Thus, AR plays an important but secondary role in strengthening safety-related decision-making among pedestrians. Yet, unlike findings from high-risk health behaviours or driver-compliance studies where AR demonstrates stronger predictive power, its more modest effect in this study may reflect cultural norms in Ghana and other developing countries that normalise at-grade crossings despite known risks, thereby reducing the emotional salience of regret.

Finally, the study establishes that CEV moderates the relationships of both SME and AR with bridge utilisation. That is, when pedestrians observe consistent enforcement of road safety rules, the influence of both social media campaigns and AR on bridge use is amplified. This supports evidence from Poulter and McKenna (2010) and Elliott et al. (2007), who argue that visible enforcement measures reinforce behavioural interventions by providing external accountability and social pressure. In this context, CEV acts as a catalyst, ensuring that the cognitive and emotional influences of SME and AR translate into consistent pedestrian safety practices. In sum, this study demonstrates that PBU is shaped by both cognitive engagement through social media and affective AR, with CEV further amplifying these effects. This reinforces the importance of integrating digital safety campaigns with institutional enforcement measures to promote sustainable pedestrian safety behaviour. This moderating effect aligns with observations in other regions where visible policing or community wardens strengthen compliance; however, the Ghanaian context, with generally low enforcement visibility, suggests that even modest improvements in enforcement can significantly enhance behavioural outcomes.

This study advances the literature in several important ways. First, by situating the analysis within the socio-cultural and infrastructural realities of a rapidly urbanising West African context, it offers comparative insights that enrich global evidence on pedestrian safety, complementing findings from health-risk behaviours, driver-compliance research and broader road-safety studies. Second, it is among the first empirical investigations to examine whether SME predicts actual pedestrian-bridge utilisation, moving beyond the awareness-focused metrics that dominate much of digital road-safety research. Third, it introduces AR as a mediating cognitive–emotional mechanism linking digital engagement to real-world safety behaviour within a pedestrian context, an area that has received limited empirical attention. Fourth, it incorporates CEV as a moderating contextual factor, providing a more integrated understanding of how digital cues, psychological processes, and institutional environments jointly shape safe-crossing decisions. Collectively, these contributions underscore the unique value of the study and offer theoretical and practical insights for designing more effective digital road-safety interventions in rapidly urbanising settings.

This study makes several theoretical contributions. First, it extends PMT by demonstrating its applicability to pedestrian safety behaviour, an area less studied compared to health or driver-focused contexts. By framing SME as threat appraisal, AR as coping appraisal, and CEV as an extrinsic reinforcement mechanism, the study broadens PMT by integrating digital and institutional dimensions. Second, it advances behavioural theory by highlighting the combined role of cognitive (SME) and affective (AR) pathways in shaping safety behaviour. Unlike prior research that examined these elements separately, the study shows their interaction in explaining PBU. Third, the results reveal that cognitive engagement exerts a stronger effect than affective mechanisms, suggesting a hierarchy of influences that refines behavioural models. Finally, incorporating enforcement visibility as a moderator enhances theoretical understanding of how institutional and community-level factors reinforce individual decision-making for sustainable behavioural change.

The findings of this study also carry several practical implications for policymakers, transport authorities and road safety campaign designers. First, the strong effect of SME on PBU suggests that digital platforms should be prioritised as a central tool in road safety campaigns. More specifically, transport authorities can deploy SME tools such as targeted Facebook and Instagram safety ads, short-form TikTok videos demonstrating safe crossing practices, WhatsApp community broadcasts for micro-targeted safety reminders, and X/Twitter alerts during peak traffic hours. These tools can be used to deliver timely, location-specific safety guidance, reinforce ongoing enforcement campaigns and sustain pedestrian awareness by providing repeated visual and interactive messages across the platforms most frequently used by urban commuters. These tools allow agencies to reach high-risk groups, including youth, market users and commuters, with context-specific, timely messages tailored to local mobility patterns. Authorities can leverage popular social media channels to deliver targeted, interactive, and emotionally engaging content that resonates with pedestrians, especially younger populations who are heavy users of digital media. Examples include user-generated content challenges promoting bridge use, influencer partnerships that normalise safe crossing and location-based geotargeted posts near high-risk corridors. Second, the role of AR highlights the importance of designing campaigns that not only inform but also evoke emotional reflection, for instance, by illustrating the potential consequences of unsafe crossing. Such affective messaging can complement informational appeals to strengthen behavioural intentions. Third, the moderating role of CEV underscores that communication efforts alone are insufficient without corresponding institutional reinforcement. Visible enforcement measures such as the presence of traffic wardens, CCTV monitoring or community policing can amplify the effectiveness of digital campaigns by providing social accountability and reinforcing desired behaviours. Together, these implications suggest that a hybrid strategy combining digital engagement, emotional appeals and visible enforcement is likely to be most effective in promoting sustainable pedestrian safety practices.

This study offers important implications for policymakers, transport authorities and road safety managers. Social media should be strategically positioned as a central communication platform, with campaigns designed to be interactive, relatable and shareable to maximise reach among digitally active groups. Beyond information sharing, campaigns should evoke emotional reflection, particularly AR, through survivor stories, testimonies and visual simulations that highlight the dangers of unsafe crossing. However, digital interventions alone are insufficient without visible community enforcement. The consistent presence of wardens, community policing, CCTV monitoring, and clear signage around bridges enhances accountability and reinforces compliance. A hybrid strategy combining digital engagement with enforcement is therefore vital for long-term behavioural change. Interventions should also target high-risk groups such as youth, informal workers and the elderly by tailoring content to local languages and cultural contexts for stronger relevance and adoption. Finally, continuous monitoring and evaluation using data analytics will enable refinement of campaigns and optimisation of resources. Collectively, these measures form an integrated framework that can improve bridge utilisation and promote safer urban mobility. In relation to the United Nations SDG 3.6, which aims to halve global road traffic deaths and injuries by 2030, this integrated approach highlights that progress requires more than infrastructural upgrades. By combining SME tools, emotional motivators such as AR, and visible enforcement mechanisms, policymakers can design interventions that meaningfully complement global efforts to reduce pedestrian injuries and advance safer mobility outcomes.

The study's findings that SME is the strongest predictor of PBU, AR has a modest effect, and CEV significantly strengthens both highlight challenges in achieving sustainable outcomes. While digital campaigns effectively shape awareness and behaviour, their impact may diminish if messages become repetitive or if structural barriers, such as poorly designed or inaccessible bridges, persist. Without complementary infrastructural improvements, many pedestrians may still prioritise convenience over safety. At the policy level, the moderating effect of enforcement visibility underscores the limitations of relying mainly on infrastructure provision while neglecting behavioural reinforcement and consistent enforcement. Without integrated frameworks that combine digital engagement, emotional appeals and visible enforcement, compliance is likely to remain low, particularly among high-risk groups such as youth, informal workers and the elderly. Managerially, limited resources, fragmented responsibilities and weak monitoring systems also constrain sustainability. These challenges can be mitigated by hybrid strategies that pair digital campaigns with visible enforcement and the development of inclusive, user-friendly infrastructure. Policies should integrate digital innovation with enforcement mechanisms and align with socio-cultural contexts, while managers can improve outcomes through inter-agency collaboration, efficient resource use and data-driven monitoring. Overall, SME, reinforced by enforcement visibility, provides the most effective pathway to improving bridge utilisation and advancing safer urban mobility.

Despite its contributions, this study has some limitations. First, the use of self-reported measures may introduce social desirability bias or recall errors. Future research could complement survey data with observational or experimental methods to provide more objective assessments of pedestrian behaviour. Second, the study examined PBU as the primary behavioural outcome. While informative, pedestrian safety is multidimensional and also involves behaviours such as jaywalking, crossing at undesignated points and adherence to traffic signals. Broadening the scope of behavioural outcomes would offer a more comprehensive understanding of pedestrian safety. Third, the study focused on SME, AR, and enforcement visibility, yet other psychological and contextual factors such as risk perception, peer influence, cultural norms or infrastructure quality may also play significant roles. Future studies should consider integrating these variables into expanded theoretical models. Nonetheless, the findings provide a strong foundation for advancing the integration of cognitive, affective and institutional mechanisms in explaining pedestrian safety behaviour. Additionally, the study's findings should be interpreted within the cultural and geographical context of major urban centres in Ghana. Because cultural norms, enforcement practices, and pedestrian behaviours differ across countries and even across regions within Africa, the behavioural mechanisms identified, such as the interplay between AR, SME, and perceived community enforcement, may manifest differently elsewhere. Consequently, the generalisability of the findings may be context-bound. To strengthen external validity, future research is encouraged to replicate and test the model in diverse cultural and geographical settings, as such comparative investigations would help determine which behavioural patterns are universal and which are shaped by local cultural influences.

Furthermore, the study affirms that the study was conducted in strict accordance with ethical principles for research involving human participants. All procedures complied with both institutional review requirements and international ethical standards, including the Declaration of Helsinki. Appropriate safeguards were implemented to ensure respect for participants' rights, dignity, confidentiality, and voluntary participation.

Prior to data collection, participants were given comprehensive information about the objectives of the research, the study procedures, and their rights as respondents, including the freedom to decline participation or withdraw at any stage without penalty. They were also assured of the confidentiality and anonymity of their responses. Informed consent was then secured in writing from all participants before they were enrolled in the study.