2.1 Conceptual foundations and theoretical lenses

Risk perception refers to individuals’ subjective judgment regarding the likelihood of and the severity resulting from a work accident (Gierlach et al., 2010). In occupational settings, and particularly in high-risk industries like construction, risk perception plays a critical role in shaping safety-related behaviors and decisions. Risk perceptions are not merely a reflection of risks’ objective characteristics but a cognitive process influenced by personal experience, social environment, and cultural context.

Building on the Theory of Planned Behavior (TPB) (Ajzen, 2005), risk perception is conceptualized as a key antecedent of safety intentions and subsequent behavior, influenced by personal attitudes, subjective norms, and perceived behavioral control (Man et al., 2021b). The Protection Motivation Theory (PMT) expands on this framework by distinguishing between threat appraisal-comprising perceived severity and vulnerability-and coping appraisal, which reflects perceived response efficacy and self-efficacy (Taylor & Snyder, 2017; Man et al., 2021a). PMT has proven particularly useful in explaining why some workers choose to engage in protective behaviors while others do not.

From a more socially embedded perspective, the Social Identity Theory suggests that individuals’ risk perceptions are shaped by their identification with specific workgroups (Andersen et al., 2015). In this view, social norms and collective practices within the work environment significantly influence how hazards are perceived and acted upon. This perspective is particularly salient in construction, where tasks are often carried out in teams under peer supervision rather than direct managerial oversight.

Finally, the Dual-Pathway Model of Risk Perception integrates affective and cognitive components, positing that emotional responses to risk (e.g., fear, worry) can operate independently of rational probability assessments and might even override them under certain conditions (Gierlach et al., 2010). This framework has been influential in recognizing the role of emotional intelligence and situational awareness in occupational safety contexts.

Taken together, these theoretical approaches highlight that risk perception is not a fixed personal trait but a multidimensional process shaped by individual, social, organizational, and cultural factors. In what follows, based on insights linked to the above-mentioned theoretical lenses, I present an overview of the determinants of workers’ risk perceptions.

2.2 Determinants of construction workers’ risk perception

A substantial body of empirical research has investigated the main determinants of workers’ risk perception (e.g., Caponecchia & Sheils, 2011; Rodríguez-Garzón et al., 2016; Ricci et al., 2019; Xia et al., 2020; Lee et al., 2021). These factors can be grouped into different interrelated domains; however, for the purposes of this research, in this study I focus on the role over risk perception of variables related to personal characteristics, work stress, coworkers’ behavior, and cultural or regional-specific factors.

Labor experience and demographic characteristics.-Early studies examined the influence of various individual profile characteristics on risk perception, including age, marital status, and labor experience. While intuitive, the reported findings have been inconsistent. For example, Caponecchia and Sheils (2011) observed that older and more experienced workers tend to perceive less risk than their younger counterparts, an effect attributed to the greater level overconfidence in the former group as a result of their experience. In contrast, younger or less experienced workers often demonstrate heightened sensitivity to work risks and hazards, particularly in their first years in the industry. However, these trends are not uniform across research analyses. Several studies have indicated that individual personality traits, such as cognitive reflection, might moderate the relationship between experience and perceived risk (Ricci et al., 2019). Furthermore, socio-demographic factors such as marital status and educational attainment also show mixed effects. Specifically, evidence suggests that married workers may express greater risk awareness, potentially due to family responsibilities, although these associations tend to dilute once workplace and group-level variables are accounted for (Caponecchia & Sheils, 2011).

Coworkers’ behavior.-Social influence-especially from peers-has emerged as a crucial factor shaping individuals’ risk perception. Unlike manufacturing environments where supervision is often direct and formal, construction workers primarily operate semi-autonomously, relying heavily on peer coordination and informal norms (Lafuente et al., 2018). Xia et al. (2020) demonstrate that coworkers’ attitudes and behaviors regarding safety significantly impact how individual workers assess work risks. For example, when peers regularly wear personal protective equipment (PPE), follow safety procedures, and speak openly about work risks, others are more likely to interpret risks as real and imminent; whereas when risky behaviors go unchallenged they become normalized, leading to a collective underestimation of danger.

This social modeling effect is underpinned by both observational learning and the desire for group acceptance. Workers often calibrate their risk perceptions based on what is seen as ‘acceptable’ within the group (Andersen et al., 2015). Arguments rooted in the theory of planned behavior support this mechanism, positing that subjective norms-i.e., beliefs about what others expect-directly shape behavioral intentions. In the construction industry, this means that if coworkers signal indifference to safety, others might internalize similar perceptions, regardless of their formal training or personal beliefs (Xia et al., 2020). Leadership also plays a role in shaping peer behavior. Studies have shown that when supervisors foster a positive safety climate-by recognizing safe acts, encouraging reporting, and intervening consistently-peer norms tend to align with safety objectives. However, in environments with weak or inconsistent supervision, peer norms often take precedence, for better or worse (Choi et al., 2019). These dynamics underscore the importance of leveraging coworkers’ influence in safety interventions and recognizing the group as a critical unit of risk perception formation.

Work stress.-Occupational stress can distort risk perceptions. Stress is induced by factors such as time pressure, heavy workload, lack of control, and poor organizational support (e.g., Gómez-Bull et al., 2023; Mastrantonio et al., 2024). When demands overwhelm available coping resources, stress levels rise, leading to diminished situational awareness, tunnel vision, and impaired hazard recognition (Zong et al., 2025). Research indicates that high stress levels impair safety behavior and can reduce perceived risk, for example, because workers tend to normalize unsafe conditions to manage emotional load. Additionally, psychosocial safety climate-workers’ collective belief that their psychological wellbeing is valued-can buffer these effects, reinforcing accurate risk interpretation even under stress (Zong et al., 2025).

Cultural or territorial differences.-Cultural frameworks and regional differences shape how individuals interpret risk messages and respond to them. Martínez-Fiestas et al. (2017) show that in collectivist cultures, workers might perceive higher social costs for deviating from group norms, even when their own safety is at stake. In contrast, more individualistic contexts might encourage personal risk assessment but also produce fragmented safety cultures. Within countries, regional variations in enforcement, training availability, and labor-market structures further modulate risk perceptions. Ricci et al. (2019) observed that migrant or ethnic minority workers often exhibit distinct patterns of risk appraisal, reflecting both language barriers and differing prior experiences with regulatory systems. These contextual factors interact with personal and social variables to produce nuanced, often region-specific, configurations of risk perceptions. In Costa Rica, contrasts between regions such as Cartago and Limón offer fertile ground for exploring how differing institutional environments and work cultures shape safety-related attitudes.

In summary, risk perception among construction workers is a complex, multidimensional process influenced by a variety of factors (e.g., individual profile, work stress, coworkers’ behavior, and cultural o territorial differences). Understanding these dynamics is essential for developing context-sensitive interventions that address not only individual awareness but also the collective practices that determine how risks are perceived and managed on the ground. The theory and evidence presented in this section provides the foundation for the present empirical investigation, which explores how coworkers’ behavior, work stress, and regional setting jointly influence risk perception among Costa Rican construction workers.

Table 4 reports the results of the OLS specifications. The first column shows the findings when the higher-order construct (i.e., ‘risk perception’) is the dependent variable, whereas the remaining three specifications test the effect of the explanatory variables over the three risk dimensions analyzed in the study, namely ‘falls’, ‘objects’, and ‘machines and equipment’. Notice that variance inflation factor (VIF) values were computed for all explanatory variables, and the average VIF is 1.41 (range: 1.03-1.77). This result confirms that the estimations in Table 4 are not contaminated by multicollinearity.

Concerning the key findings of the study, from Table 4 it can be seen that, for all model specifications, the dummy variable capturing the effect of co-workers’ safety behavior is positive and statistically significant. This result suggests that workers tend to follow social norms within their work environment and, for better or worse, they routinely imitate their peers’ safety behaviors (Andersen et al., 2015; Xia et al., 2020). The stability of the coefficient across the three risk groups indicates that the peer signal is not risk-specific but generalizes to the overall safety environment on site.

In addition, contrary to prior work drawn from the job-demands literature (e.g., Gómez-Bull et al., 2023; Mastrantonio et al., 2024), in all model specifications the dummy variable linked to work stress is negative but not significant. One interpretation for the lack of significance is that the relationship between work stress and perceived risk might be more complex than initially expected. For example, the literature indicates that stress can impair situational awareness (in this case, on site) and reduce the ability to evaluate risks accurately (Taylor & Snyder, 2017). Related, stress might reduce or minimize engagement with safety behaviors, but only in cases where the work environment is plagued with poor safety practices or lack supervisory support (Lee, 2022).

Finally, the results show that the variable linked to the regional effect (Limón) is negative and highly statistically significant in all models. The findings for the coefficients indicate that, for the analyzed risk scenarios (Appendix 1), workers in Limón perceive substantially less risk than their counterparts in Cartago. The result corroborates descriptive evidence of lower means and distinct distributions for all risk perception constructs (the three first-order and the higher-order construct) (Table 2), and underscores the influence of contextual factors such as enforcement intensity, subcontracting chains, and local work culture on perceived risk (Ricci et al., 2019).