This study aimed to measure individuals’ attitudes towards presenteeism and test the suitability of the TPB for explaining presenteeism behaviour. The cluster analysis of the developed vignettes to measure the individuals’ attitude toward presenteeism resulted in a satisfactory three-component solution. The three components can be clearly distinguished, and the validation by applying PCA and the Jaccard Index resulted in a good fit.

The three clusters show clear differences regarding presenteeism behaviour. Individuals assigned to cluster A show a rather positive attitude towards presenteeism behaviour across all vignettes compared to the other clusters. Individuals assigned to cluster B show a rather negative attitude towards presenteeism behaviour and decide to call in sick when they perceive common symptoms. Individuals assigned to cluster C somehow decide to call in sick when they have an objective indicator, such as a temperature of 38.5 °C (vignettes 2–3), and without such objective indicator, a positive attitude towards presenteeism behaviour.

This pattern could be explained by the different lay concepts of health by Faltermaier (1994). He described three forms of health behaviour, which could be assigned to the above-described clusters. Cluster A is characterised by action and performance-oriented health awareness. Individuals in this health concept ignore health complaints until their ability to work is impaired. The importance of health and health maintenance is low, and work is a priority (Faltermaier 2023). Cluster B is described by the health concept of psychologically influenced and multidimensional health awareness. Individual health is a high priority among individuals, and they strive for well-being, peace, balance, and the absence of pain and discomfort. The individuals have high health awareness, and symptoms are understood as signals of overload and psychological conflicts (Faltermaier 2023). Cluster C may be characterised by the health concept organic-medical health awareness. Health is understood as the absence of illness and illness is perceived if identifiable symptoms are present. The importance of health is also low in this group (Faltermaier 2023).

However, this comparison is purely hypothetical and requires further detailed analysis to determine whether the theoretical framing of the identified cluster is correct. Thus, it would be necessary to identify the characteristics of the three groups. For example, the theory assumes that women, in particular, have a higher level of health awareness and are, therefore, more likely to be assigned to the health concept of psychologically influenced and multidimensional health awareness (Faltermaier 1994). In contrast, the literature shows a rather heterogeneous picture of presenteeism behaviour differences between females and males (Webster et al. 2019). Regarding presenteeism, the profession and the setting seem more decisive, such as social professions (Aronsson et al. 2000; Chambers et al. 2017), which naturally have a disproportionate number of women.

We were able to evaluate the suitability of the TPB on presenteeism behaviour. As the theory explains, attitude, subjective norm, and perceived behavioural control are associated with the behaviour. Our model explains 29.6% of the variance in the fixed effects and 51.6% of the variance in the fixed and random effects combined and the higher value compared to the marginal R² suggests that a significant portion of the variability in presenteeism is also due to individual differences captured by the random effects (subjects).

The individual’s attitude seems to be a highly relevant factor for presenteeism behaviour. As discussed in the paragraph above, the individuals’ health awareness may form the attitude. This awareness is mutually shaped by lifestyle. However, it is not static and can change due to dealing with health issues (Faltermaier 2023).

Subjective norms, such as workplace expectations or peer behaviours, significantly influence whether employees attend work while ill. This insight underlines the importance of communicating about illness in the workplace. A health culture in the company that promotes a discussion about one’s health appears to be a potentially effective basis for reducing presenteeism.

Management plays a key role in the formation of the presenteeism climate. They have, in addition, a role model, and research showed that their presenteeism behaviour influences the employees’ behaviour (Dietz et al. 2020). The results of the mixed effects regression show that management’s handling of illness is linked to presenteeism. The feeling of being able to talk to direct superiors about illness was found to be associated with lower presenteeism. This finding aligns with other research on leadership and presenteeism behaviour (Ferreira et al. 2015), particularly leadership styles such as transformational (Cicek and Kilinc 2020), inclusive (Qian and Wang 2023) or health-oriented (Vonderlin et al. 2021). Inclusive leadership is characterised by openness and aims to build psychological safety (Cicek and Kilinc 2020); it may provide a fruitful ground for establishing a positive health culture at work. However, there also appears to be a risk of increased presenteeism behaviour due to such management styles under certain conditions, such as increased subjective obligation (Qian and Wang 2023).

Perceived behavioural control, reflecting factors like job insecurity and workload, also played a key role, affecting employees’ perceived ability to make autonomous decisions regarding their health. Higher perceived job insecurity and quantitative demands were associated with a higher presenteeism behaviour.

The findings underscore the importance of addressing both social and organizational expectations (subjective norms) and workplace demands (perceived behavioural control) in efforts to mitigate presenteeism. By focusing on both aspects, organizations can create environments that support healthier decision-making, allowing employees to stay home when ill without fearing negative consequences.

Measures are therefore needed to enable leaders to communicate with their employees about the subjectiveness of illness. They must be aware of the adverse consequences that proposed leadership styles can have. In this context, research is needed to determine how open communication and a health-promoting culture at work can be implemented without achieving the opposite results in presenteeism through increased commitment and subjective obligation. One effective approach to improving workplace culture and addressing presenteeism is implementing health-oriented leadership interventions. Health-oriented leadership emphasizes the role of supervisors in promoting health by focusing on self-care and staff care. A specific example is mindfulness- and skill-based health-oriented leadership intervention developed by Vonderlin et al. (2021), which demonstrated significant improvements in supervisors’ mental health and health-promoting behaviours. Regarding a healthy work culture, research suggests that improving working conditions is related to improved presenteeism. A recent systematic review concluded that only multicomponent programs on different organizational levels positively affected presenteeism (Støren and Grønningsæter 2024). Such successful multi-component programs comprised individual health and lifestyle. In addition, applying participatory processes for the program implementation was identified as favourable, which in turn is associated with the leader’s behaviour.

This article follows the STROBE reporting guidelines for longitudinal studies (Von Elm et al. 2007). The study also benefits from a large sample size, facilitating robust statistical analyses and providing deeper insights into our research questions. This extensive sample size improved our ability to accurately identify associations within the longitudinal data and reduced the likelihood of random results, thereby enhancing the study’s internal validity. Additionally, the study’s strength lies in its use of valid and reliable scales. The longitudinal design offers distinct advantages in understanding presenteeism behaviour over time. Unlike cross-sectional studies, which provide a snapshot of behaviour at a single point, our longitudinal approach allows us to observe changes and trends in presenteeism behaviour, providing deeper insights into how attitudes, subjective norms, and perceived behavioural control evolve and influence presenteeism over time.

One limitation of the study is that the complete TPB could not be applied, as no intention was measured, but rather presenteeism behaviour directly. We can, therefore, only make a statement about the extent to which the individual factors are directly related to the behaviour, but not which intention they generate, which ultimately leads to the behaviour. While interaction effects, such as those between subjective norm and perceived behavioural control, may offer additional insights into the dynamics of presenteeism, these were not included in the present analysis. Our primary aim was to evaluate the direct contributions of attitude, subjective norm and perceived behavioural control on presenteeism. Future research could explore these interaction effects further to refine the understanding of the model in this context.

The vignettes developed also only describe physical symptoms of common illnesses. Psychological impairments, such as depressive or burnout symptoms, are not currently included. However, mental health is found to be equally relevant in the context of presenteeism behaviour (Ruhle et al. 2019). Future research should expand the vignettes to include this perspective, particularly about common psychological impairments among the working population, such as burnout symptoms (Van Der Molen et al. 2020). The addition may contribute to an improvement of the identified clusters.

Furthermore, the voluntary nature of the questionnaire may have led to the underrepresentation of specific employee groups. This could result in biased outcomes, as the perspectives of these underrepresented groups might differ significantly from those who chose to participate. Consequently, the findings may not fully capture the diversity of experiences and opinions within the entire workforce, potentially limiting the generalizability of the results. However, the differences identified in the cluster analysis indicate that, despite the survey method, a sufficiently broad range of participants with different attitudes towards presenteeism could be included. Additionally, the response rate decreased from 40 to 20% between the measurement points due to companies who decided not to participate in a second round. Reasons for dropouts of the companies were too many inquiries, such as from research and annual job satisfaction surveys and the appointment of a new manager, resulting in a shift in priorities. It is also important to note that the sampling was drawn primarily from specific industries with different proportions, which may limit the generalizability of the findings across other sectors. The workplace culture, job demands, and presenteeism patterns in other industries differ (Marklund et al. 2021), and future research could address this by including a broader range of sectors to enhance the applicability of the findings.