Working sick and out of sorts: a cross-cultural approach on presenteeism climate, organizational justice and work–family conflict*

Abstract A climate of presenteeism has important effects on employee well-being and the organization itself. Our study, based on surveys of health sector employees in six different countries (Brazil, Ecuador, Lebanon, Portugal, Russia and Spain) examines whether organizational justice plays a mediating role in the relationship between a presenteeism climate in the organization and work–family conflict (WFC). Our results indicate that the perception of organizational justice and the presenteeism climate do influence WFC. Moreover, higher levels of WFC were found in non-Latin countries. This study contributes to the work attendance and life balance field by providing cross-cultural empirical evidence corroborating the effect of justice and presenteeism climate on the WFC.


Introduction
The time employees spend at work and the impact this has on related work-family conflict (WFC) is of interest to both researchers and practitioners. Presenteeism is "the problem of workers being on the job but not fully functioning, because of illness or other medical conditions (physical or psychological)" (Hemp, 2004: 49). This phenomenon is more prevalent in companies with high presenteeism climate where both the co-workers and supervisors may pressure employees to continue working beyond the time necessary for efficient performance at work. This is the result of competitiveness, extra-time valuation, difficulty of replacement and lack of supervisor support (Ferreira, Martinez, Cooper, & Gui, 2015;Zhou, Martinez, Ferreira, & Rodrigues, 2016). Prior research has established that working time arrangements are important determinants of presenteeism in organizations (Bockerman & Laukkanen, 2010). According to the equity theory (Adams, 1965), these effects may be mediated by any feelings of injustice due to perceptions of unequal treatment at work. From a comparative perspective, the relationship between these factors may vary across samples from different national cultures (Stock, Strecker, & Bieling, 2016). This paper thus explores the interrelationship between these issues, using empirical data gathered from healthcare professionals in six countries. We aim to clarify the relationship between a presenteeism climate and WFC and the role of cultural differences in this relationship. The evidence indicates that family is a more extensive and pervasive aspect in Latin countries (House, Hanges, Javidan, Dorfman, & Gupta, 2004) and, as these are under-studied countries in presenteeism research, we concentrate on those countries. To check our findings, we compare them with evidence from the very different contexts of Russia and Lebanon.
Despite the growing interest in WFC among academicians and practitioners, most studies have been conducted with US samples (Byron, 2005). Casper, Allen, and Poelmans (2014) edited a special issue in the Journal of Applied Psychology on WFC and confirmed the lack of studies on the work-family interface outside North America and suggested a research agenda of culture influences on the work-family domain. Prior studies reported differences in WFC and interfaces in countries with higher levels of gender equality (Behan, Drobnič, & Präg, 2014) and also in individualism-collectivism cultural orientation, as well as in the power distance dimension (Billing et al., 2014). We used a cross-cultural design to compare Latin and two contrasting non-Latin countries that despite the differences share the high collectivism/ family orientation of the Latin countries (Hofstede, Hofstede, & Minkov, 2010). Most of the considerable amount of research on presenteeism has been published over the past decade considers developed countries, namely those in North America and Scandinavia (Bockerman & Laukkanen, 2010;Johns, 2010Johns, , 2011; as far as we have been able to discover, there are no extant studies considering differences between Latin and non-Latin countries as a cross cultural variable explaining WFC differences. Thus, we aim to advance the presenteeism literature by answering calls from previous studies for further exploration of presenteeism based on a cross-cultural approach . Meta-analytic results have also shown that work characteristics, role stressors, social support and personality are WFC antecedents (Michel, Kotrba, Mitchelson, Clark, & Baltes, 2011).
Performing hard at work might reduce the energy and personal resources for other life domains such as the family (Hangis, Kotrba, Zhdanova, & Baltes, 2011). Also, despite previous research that links the work schedule framework and schedule flexibility practices as antecedents to presenteeism behaviour (Johns, 2010), to our knowledge no studies have explored the presenteeism climate as an antecedent variable to WFC. A presenteeism climate results from beliefs and values about the sector, department or organization and also the society (Nicholson & Johns, 1985) that pressure employees to attend work despite being ill. In other words, a presenteeism climate is characterized by individual aspects of each employee, (i.e., personal values of reduced perception of the legitimacy of absenteeism), characteristics of the job (i.e., responsibilities and task interdependence) and contextual characteristics (i.e., competitions among workers and pressure from supervisors) (Ferreira et al., 2015;Zhou et al., 2016). A presenteeism climate is contingent on values or beliefs embedded in the society, thus further suggesting the importance of research into how cultural differences affect the influence of presenteeism climate on employees' behaviour.
WFC has significant correlations not only with presenteeism, but also with organizational justice and stress (Judge & Colquitt, 2004). The organizational justice construct concerns the degree to which people feel that they have been treated fairly by their organizations and those in their immediate surroundings (e.g., customers or patients) (Smith, Bond, & Kağitçibaşi, 2006). It has been argued that globalization has raised the visibility of justice within organizations (Fischer & Smith, 2003). However, few researchers have studied the influence of cultural values on organizational justice (Shao, Rupp, Skarlicki, & Jones, 2013). For example, Latin countries tend to be characterized by charismatic leaders and supervision that promotes a potentially collective cultural dimension (House et al., 2004). On the contrary, in non-Latin countries restraint cultures tend to hide the role of charismatic leaders. Simultaneously, perceived organizational support (House et al., 2004) makes employees more dependent on other variables such as distributive justice mechanisms to mitigate the effects of presenteeism climates on WFC. Previous studies showed that contrary to non-Latin countries (such as India), in Latin countries (Spain and Peru) higher levels of management support lead to lower scores on work life conflict (Agarwala, Arizkuren-Eleta, Del Castillo, Muñiz-Ferrer, & Gartzia, 2014). Taking into account these cultural differences, we believe that our research enriches conceptualization of presenteeism by including it and its effects in the WFC domain and testing it with samples from different countries. Further studies will help us to understand whether the cultural correlates of work-family issues specific to the countries mentioned (Agarwala et al., 2014) or are more generalizable. Our study examines whether differences exist in these relationships in Latin countries, contrasting them with two non-Latin countries with high collectivism, family orientation and low levels of indulgence (Hofstede et al., 2010). This paper starts by outlining the literature in these fields and from that develops a model and our hypotheses. Next, we explain the research design and methodology and present our empirical findings. Finally we discuss our results and draw conclusions for practitioners and for further research.

A climate of presenteeism
Presenteeism is an emerging concept in the organizational behaviour domain (Johns, 2010).
According to Halbesleben, Whitman, and Crawford (2014), presenteeism and absenteeism have developed along different paths, even though they are tied into a single decision (to stay at home or to go to work). In fact, the decision to go or to stay at home may be related to the kind of relationship employees and their supervisors have. Both presenteeism and absenteeism are strategies that employees adopt to deal with dialectical tensions between supervisors and subordinates, among other factors (Halbesleben et al., 2014).
There is still no consensus about how to measure the presenteeism construct and its dimensions (Cooper & Lu, 2016;Johns, 2010). For example, some scholars (e.g., Simpson, 1998) emphasize concepts such as a 'presenteeism climate' at a theoretical level, as there is a lack of psychometric instruments to test individuals and a need for a more in-depth rooted conceptualization Gosselin, Lemyre, & Corneil, 2013).
The financial crisis that began in 2008 led to some organizations downsizing or even closing. In this context of job insecurity, presenteeism behaviour tends to increase .
Simultaneously, companies are continuously seeking cost efficiency (Simpson, 1998) by stimulating internal competition, again creating a climate of presenteeism. This will differ between countries and contexts. Job demands, social pressure, and job insecurity also contribute to presenteeism (Johns, 2010) as do injustice, uncertainty and fear (Aronsson, Gustafsson, & Dallner, 2000;Halbesleben et al., 2014).
To be present at work does not, of course, guarantee productivity. People going to work while suffering health problems may reduce their performance and that of those around them (see Johns, 2010, for a review). Resultant losses may be more serious than that caused by employee absenteeism (Hemp, 2004;Hummer, Sherman & Quinn, 2002).

Work/ family conflict
WFC has been defined as a "specific form of inter-role conflict in which work and family roles are mutually discordant in some respect" (Allen, Cho, & Meier, 2014: 101). Research has primarily focused on the organization-level. We seek to extend previous findings by focusing our research on the individual and national levels and examining some underlying cultural issues.
Work-family conflict has been identified as a two-dimensional construct where work interferes with family and family interferes with work (Frone, Yardley, & Markel, 1997). WFC has different manifestations across countries due to the material and psychological interdependence of the differing family models. Material interdependence appears stronger in collectivistic societies, whereas psychological interdependence has no cultural specificity (Smith et al., 2006). Collectivist societies are those cultures oriented to groups (e.g., Eastern Asian and Latin countries). In fact, most studies comparing individualist and collectivist cultures involve comparisons between the USA and Eastern Asian countries rather than Latin countries (Torres, 2009).
The increasing interest in non-traditional gender roles, working hours and dual-income households has also been the subject of WFC research (Clark & Weismantle, 2003). Meta-analytic results have demonstrated that work schedules, work role stressors, work social support, work characteristics and personality are antecedents to work/ family conflict (Michel et al., 2011).
Strenuous physical or emotional commitment at work can also reduce the resources available for other life domains such as family (Hangis et al., 2011). These kinds of situation-based variables are important antecedents to WFC.
Empirical research provides evidence that country characteristics play a significant role in the loss of productivity due to presenteeism (Knies, Candel, Boonen, Evers, Ament, & Severens, 2012). Other studies have found cross-cultural differences for mental and physical presenteeism between Indian and US employees (Garczynski, Waldrop, Rupprecht, & Grawitch, 2013) and also between British and Chinese workers . In fact, the way people cognitively organize their daily lives varies from culture to culture.

Distributive justice and cultural values
WFC has significant correlations not only with presenteeism but also with organizational justice and stress (Judge & Colquitt, 2004). The organizational justice perception has to do with the degree to which people determine that they have been fairly treated by their organization and by those, such as customers or beneficiaries, in their immediate surroundings (Smith et al., 2006). Hart and Cooper (2001) found that aversive emotional and psychological work experiences are related to perceived unfairness and uncertainty, affecting employees' WFC. Moreover, this perception is higher in countries with low uncertainty avoidance, because employees have less control over the systems and procedures that could ensure fairness (Hofstede, 1993).
How people distribute or allocate resources has been of interest to scholars in the past (Greenberg, 1990) and is increasingly being recognized as important (Fischer & Smith, 2003).
However, few researchers have studied the influence of cultural values such as uncertainty avoidance on organizational justice (Shao et al., 2013). This gap is noteworthy, as high levels of uncertainty avoidance are partly defined by high levels of intolerance towards deviant behaviour (Hofstede, 1980), thus affecting perceptions of justice.
Cultural values can be summarized as commonly-held standards of what is acceptable, important, right or workable in any community or society (Rokeach, 1973). One of the most used models of cross-cultural comparison is the GLOBE (Global Leadership and Organizational Behaviour Effectiveness Research) project. This project has its roots in previous cross-cultural research findings (Hofstede, 1980) but it specifically identifies different cultural competencies and groups countries into societal clusters (House et al., 2004).
The GLOBE model includes the clusters 'Latin European' and 'Latin American' (House et al., 2004). The two clusters are linked by the colonial history and languages of Latin America, having similar roots and sharing much in common (Kogut & Singh, 1988;Laurent, 1983). Latin clusters tend to have a high performance orientation, representing the need to seek high standards in decision-making and innovation. This is a charismatic value-based dimension where people are inspired and their passion is the engine that leads to performance. The collective dimension plays a key role in encouraging and rewarding group members to perform well (House et al., 2004).
Collectivism and leaders' charisma ensure employee's confidence and stability and also influence worker perceptions of organizational justice.
There is evidence that each of the issues related to presenteeism is influenced by national cultural values . Casper et al. summarized studies of the work-family domain and called for further research on the context of cultural influences (Casper et al., 2014). Other studies have reported differences in WFC and interfaces in countries with higher levels of gender equality (Behan et al., 2014) and differences related to the individualism-collectivism and power distance cultural dimensions (Billing et al., 2014;Stock et al., 2016). Gender equality, individualism and power distance issues have a substantial impact in Latin countries, which tend to have high uncertainty avoidance, focusing on planning and creating stability as a way of dealing with life's uncertainties (Smith & Bond, 1999).

Research model and hypotheses development
On the basis of the above theoretical framework, we present our hypothesized model in Previous research has shown that presenteeism and WFC are related constructs (Hammer, Bauerm, & Grandey, 2003). Indeed, absence from home promotes WFC which, in turn, is found to have crossover effects on organizational withdrawal. The development of a climate of presenteeism results from the aggregation of two distinct but complementary spheres: societallevel beliefs and specific sectorial, organizational or departmental beliefs (Nicholson & Johns, 1985). Presenteeism can be seen as either a counter-productive behaviour or as a prime example of organizational citizenship behaviour, particularly in interdependent environments (Johns, 2010). The perceived legitimacy of presenteeism will depend on how far the society and the organization admit certain health conditions as acceptable reasons for being absent from work (Nicholson & Johns 1985). An extensive study conducted in nine organizations (Baker-McClearn underlying reasons for presenteeism: approach and avoidance motives. Accordingly, some employees choose to attend work while ill because they believe that they should overcome the discomfort to be loyal to their jobs, colleagues, and clients (approach motives); others force themselves to attend work because of the fear of financial loss or the backlash of social condemnation (avoidance motives) . These motives are linked to dysfunctional 'competitive presenteeism' that promotes an extreme, intensive competitive culture among employees (Simpson, 1998). From this evidence, we infer that companies with prevalent presenteeism climates promote differences between desired and actual working hours and consequently a low working time balance. Thus, Bockerman and Laukkanen (2010) found that the working time match between desired and actual working hours decreases by 8% with the prevalence of sickness presenteeism. Considering that work time demands and overtime hours promote work family interferences (Geurts, Beckers, Taris, Kompier, & Smulders, 2009), we hypothesize that:

H1. A presenteeism climate is positively related to WFC.
Drawing on equity theory (Adams, 1965) we seek to increase knowledge regarding the impact of perceived justice in the indirect relationship between a climate of presenteeism and WFC. Equity theory explains that individuals are more motivated to work if they perceive that the distribution of resources is fair to both parties. A meta-analysis of 190 studies (Cohen-Charash, & Spector, 2001), revealed that organizational practices, namely adherence to justice rules or quality of treatment and communication with employees, affects perceived justice, which in turn influences individual outcomes of performance, extra-role behaviour, counterproductive behaviour and attitudes/ emotions. Organizational practices associated with a competitive presenteeism climate may reflect unfair treatments such as supervisors requiring presence even in cases of serious illness (Simpson, 1998). When a particular situation is perceived to be unfair, it affects the person's emotions, cognitions and behaviours, leading to counterproductive withdrawal or deviant behaviours (Krischer, Penney, & Hunter, 2010). So the research suggests that one way to decrease WFC could be through promoting organizational justice in the workplace (Judge & Colquitt, 2004).
Distributive justice also influences work outcomes (Liao & Rupp, 2005;Zhou & Li, 2015) by reducing turnover intention (Konovsky & Cropanzano, 1991) and by increasing satisfaction with employee compensation and benefits (Colquitt, Conlon, Wesson, Porter, & Ng, 2001). The perception of justice helps employees to feel more engaged with the company, thanks to an increased perception of organizational support from peers and supervisors (Biswas, Varma, & Ramaswami, 2013). In some cases engaging in counterproductive withdrawal or deviant behaviours helped employees to cope with the unfair work environments that are part of a presenteeism climate (Krischer et al., 2010). Fairness perceptions (justice) have also been shown to influence absenteeism, the other side of the presenteeism coin (Lam, Schaubroeck, & Aryee, 2002), and may be related to WFC (Kossek, Colquitt, & Noe, 2001). Adherence to a presenteeism climate is conditionally dependent upon a perception of distributive justice or equity which, in turn, increases the time devoted to the organization affecting WFC. Accordingly, we hypothesize that:

H2. Distributive justice mediates the relationship between a presenteeism climate and WFC.
Cross-cultural differences impact these hypothesized effects. Justice can be considered a social construct , and some studies show that the concept of fairness varies across contexts (Lamertz, 2002). The cultural context may have an important effect on how justice is implemented: what is considered to be appropriate in one culture may not be appropriate in another (Fadil, Segrest-Purkiss, Hurley-Hanson, Knudstrup, & Stepina, 2004). In addition, cultural differences influence our perceptions and our actions in relation to justice (Folger & Cropanzano, 1998).
The GLOBE model includes clusters such as Latin Europe, including Portugal and Spain, and Latin America, including Ecuador and Brazil (House et al., 2004). In the Latin countries the collective dimension plays a key role, encouraging and rewarding group members to achieve greater performance (House et al., 2004). Collectivism and leaders' charisma reinforce employee confidence and stability and mitigate the role of organizational justice. By contrast, more restrained societies regulate the suppression of gratification with strict social norms (Hofstede et al., 2010), so countries such as Lebanon and Russia are more dependent on distributive justice in terms of explaining how a presenteeism climate relates to WFC. In such non-Latin countries, employees expect material reward for the job done and easily feel that they were not treated in a fair way (Hofstede et al., 2010): the perception of justice plays a pivotal role in explaining the WFC of employees working in highly demanding climates of presenteeism. The low levels of indulgence of these non-Latin countries (i.e., less free speech and more personal control) make employees more sensitive to distributive justice (Hofstede et al., 2010;House et al., 2004) and to the perception if the distribution of resources is fair (Adams, 1965). Perceptions of fairness and distributive justice allow employees from more restrained cultures to accept the 'rules' of presenteeism climates and to accept a more competitive environment (Simpson, 1998), which in turns results in higher WFC. Thus, we hypothesize that:

Methodological strategy
We conducted our research in the health sector, as it ranks high in terms of emotional costs, employee turnover, and pressure to attend work while ill (Aronsson et al., 2000;. In order to choose the countries for comparison, we consulted the GLOBE, which includes clusters, such as the Latin countries (House et al., 2004). The identification of a general Latin cluster (considering Latin European and Latin American countries) is widespread in the cultural literature and appears as a typical outcome of colonialism (Kogut & Singh, 1988;Laurent, 1983). Taking into account that there are countries from Europe and America that share the same roots and language, we opted to choose Portugal and Brazil, as well as Spain and Ecuador (both sharing the same cultural roots and language; Portuguese and Spanish, respectively). Russia and Lebanon, though not a cluster in themselves were taken as representing non-Latin countries and were adopted to contrast results with the Latin countries. Russia and Lebanon have the same cultural pattern of high collectivism, power distance and family orientation as typifies the Latin countries, as we are not comparing them to extreme cases. However, both cultures have very patriarchal and hierarchical societies, where there are high levels of mistrust between managers and employees (Dixon, Day, & Brewster, 2014;Sidani & Hakim, 2012). Moreover, these two countries are both low in indulgence (Indulgence = 25 for Russia, and 20 for Lebanon), where the Latin countries' scores are rather high (Hofstede et al., 2010;House et al., 2004). Restrained cultures such as Russia and Lebanon place less importance on freedom of speech and there is a greater sense of low support from peers and supervisors (Hofstede et al., 2010). Therefore, for statistical simplification, we grouped both countries together as a contrast group to the Latin countries.

Sample and procedure
Our research site was the health sector and our sample was drawn from health professionals (nurses, doctors, physicians, laboratory analysts, and administrative staff). We approached one public hospital (>1,000 employees) per country with the number of individuals responding being: Brazil (153), Ecuador (90), Lebanon (213), Portugal (135), Russia (101) and Spain (102). In Spain data were collected in two different hospitals. Each hospitals' Executive Boards and Ethical Committees approved the research at their institutions. Participation was voluntary and confidential. In general, data were collected through a paper-and-pencil questionnaire although in one small hospital in Spain a web-based survey was sent to 300 professionals, generating 50 responsescontributing to a high non-response rate (in Spain) of 83%. For the remaining countries, we had a random non-response rate of 70% for Brazil, 33.7% for Ecuador, 39.1% for Lebanon, 38.3% for Portugal and 15.8% for Russia. In the end, our final convenience sample consisted of 794 valid questionnaire responses (an average response rate of around 46%). The response rate variance is probably attributable to the senior managements' capacity to motivate potential participants. Two thirds of participants were female, which is representative of the sector; they were on average in their mid-30s; with more than a decade of experience and just less than a decade of seniority. About a fifth of the overall sample held supervisory positions.

Measures
The presenteeism climate was measured through the 9-item scale developed by Ferreira et al. (2015). This scale conceptualizes the presenteeism climate as the perceived institutional pressure to keep employees working beyond the time necessary for efficient performance at work (Simpson, 1998) and despite being ill (Aronsson, et al., 2000). This is a first-order construct related to three second-order constructs with three items each. These were chosen from three distinct presenteeism dimensions (co-worker competitiveness, extra-time valuation and the difficulty of replacement).
The co-worker competitiveness dimension captures the extent of pressure from colleagues and the perception that employees are stimulated to stay longer after the end of their normal work schedule (sample item: 'Some colleagues stay longer hours at work just for the sake of being noticed'). The extra-time valuation dimension represents the belief that people staying longer hours at work are more productive (e,g., 'I benefit from staying longer hours at work'). The difficulty replacement dimension explains the decision to go to work while ill due to a sense of responsibility and awareness that one's work cannot be easily replaced (e.g., 'I have to go to work even when ill, because I am necessary there'). Respondents indicated their agreement with each statement on a seven-point Likert-scale, ranging from 1 (strongly disagree) to 7 (strongly agree). Higher scores represent a higher level of a presenteeism climate in the organization. The average Cronbach's alpha is .83.
WFC assesses how work affects family life. We used a five-item scale developed by Netemeyer, Boles and McMurrian (1996) (to a small extent) to 5 (to a large extent). Higher scores represent a higher level of distributive justice. The average Cronbach's alpha is .89.

Controls.
We also obtained background information from respondents: gender and age, country, work hours across countries, job title and managerial position, as well as seniority (years in their current jobs) and experience (years in health sector). We used these as control variables.

Measurement model and common method variance
In order to control for potential common method variance due to self-reported measures in the study, we used Harman's single-factor test (Podsakoff, MacKenzie, Lee, & Podsakoff, 2003) to detect any possible effects. Accordingly, if there is common method variance, a single-factor confirmatory factor analysis model will provide better-fit indices, accounting for the majority of the covariance among all of the studied variables. Our results revealed that a single-factor model   (Henseler et al., 2009). Moreover, our findings also confirm the variables' discriminant validity with all of the Average Shared Variance (ASV) scores being below the AVE score (Hair et al., 2010). Table 1 about

Results
In Moreover, the participants of non-Latin countries reported having better health.

Structural invariance
In the next step we studied measurement invariance to test construct validity across the subsamples (Latin vs. non-Latin). In the previous section, we found that a general second-order hierarchical factor (the presenteeism climate) and three first-order factors (co-worker competitiveness, extra-time valuation and difficulty of replacement) was the model that best fit the data. We then carried out a Multi-Group Confirmatory Factor Analysis (MGCFA) to test the structural invariance of employees from hospitals belonging to countries with different cultural backgrounds. The MGCFA allows us to assess the measurement invariance by using the same factorial structure across different groups and to test fitted models with incremental invariance properties. We used changes in CFI (ΔCFI) values to compare nested values. As the models became more restrictive (Table 4), ΔCFI < .01, we expected that the data fit would not change considerably (Cheung & Rensvold, 2002).

Hypotheses testing
To test the mediation hypotheses we used the PROCESS bootstrap macro for SPSS developed by Hayes (2013). In addition to estimating the coefficients of the model using ordinary least squared regression-based path analytical framework, PROCESS generates estimations of the direct and indirect effects in mediation models (Hayes, 2013). We studied the relationship between the presenteeism climate and WFC as mediated by distributive justice. We conducted mediation analyses for the general sample and then for the different subsamples consisting of employees belonging to the different clusters (Latin versus non-Latin countries). As suggested by Preacher and Hayes (2008), we used 99% bootstrap confidence intervals (CI) to avoid problems related with possible asymmetric and non-normal sampling distributions related with indirect effects.
Presenteeism climate -Distributive justice -WFC. We found support for hypothesis 1 (see Table   5) which predicts the main positive effect between a presenteeism climate and WFC (c'= 0.61, p < .001). Furthermore, when we included distributive justice in the model, the positive correlation decreased slightly, indicating that distributive justice partially mediates the relationship between the presenteeism climate and WFC (Hypothesis 2). Moreover, a bias-corrected bootstrapping confidence interval for the indirect effect based on 10,000 bootstrap samples was entirely above zero [0.11 to 0.53]. Thus, the bootstrap result provides support for partial mediation via indirect effects supporting hypothesis 2. This suggests that distributive justice partially mediates the relationship between a presenteeism climate and WFC (Table 5). Table 5 about Cross-cultural cluster differences. We applied the same procedure to the subsamples of employees belonging to Latin and non-Latin countries (Table 6). Hypothesis 3 predicts the mediation as more salient in non-Latin countries. As with the general sample, we found a positive and significant link between a presenteeism climate and WFC for non-Latin countries (Lebanon and Russia).
Moreover, a presenteeism climate influenced distributive justice (path a = 0.394) which, in turn,

Discussion
The aim of our study was two-fold: to examine the relationship of a climate of presenteeism to WFC; and to examine the differences in that relationship across different cultural contexts. This study therefore contributes one of the first empirical to test the notion of a presenteeism climate and one of the first studies to examine the issue, and the relationship in, specifically, Latin countries. Consistent with our predictions, we found that a presenteeism climate is positively related to WFC and that distributive justice mediates this relationship. However, in the Latin countries at the focus of our study this relation is less obvious, thus raising some questions about the generalizability of Adams (1956) theory of inequity in social exchange. These findings contribute to the field of work/ family conflict by providing evidence that presenteeism acts as predictor variable rather than a dependent variable and as such makes a further contribution by challenging the assumptions in Johns (2010Johns ( , 2011 that the relationship runs the other way.
Furthermore, our results support the hypothesized differences across cultures.
The reasons for a climate of presenteeism may vary with organizational unit. Nowadays, and partly due to the current economic recession, including health budget cutbacks, some health employees fear losing their jobs, and this may pressure them to stay longer at work. They might also simply feel that they cannot easily be replaced. Both situational conditions encourage going to work despite being ill. This individual decision may have a direct effect over the short and midterm, not only for the unit's overall performance but also for the employees' families and their own personal wellbeing.

Theoretical contribution
We believe that this paper constitutes one of the first attempts to differentiate a specifically Latin perspective on presenteeism. We took into account the similar roots of both the American and the European Latin cultures considering that these countries are linked by the colonial history and language (Kogut & Singh, 1988;Laurent, 1983). This is an important contribution as research in HRM consistently finds a US-centric perspective that fails to consider similarities and differences across cultural clusters. Our study advances the literature by providing empirical evidence that a presenteeism climate relates to WFC. In a previous study, Johns (2011) mentioned the potential influence of WFC on presenteeism behaviour, but, to the best of our knowledge, no one has treated presenteeism (and presenteeism climate) from the opposite viewpoint, as an antecedent variable to WFC (Michel et al., 2011).
This study is also one of the few attempts to study, with empirical data, the construct 'presenteeism climate' (Ferreira et al., 2015). Particularly following the global financial crisis that began in 2008, employers and employees became more competitive than ever, directly and indirectly affecting the lives of a vast number of families. Our research sheds some light on the presenteeism and justice literatures. Consistent with previous contributions, our findings confirm that organizational justice has a direct impact on perceived WFC (Judge & Colquitt, 2004). In work contexts, looking at individuals in isolation has limited value, because the ways in which they perceive and interpret the organizational milieu will have an impact on this relation: as we predicted and our results corroborate.
Despite there being a call from previous studies to investigate how WFC varies across countries (e.g., Casper et al., 2014), the current study also extends equity theory (Adams, 1965) by explaining how distributive justice mediates employees working in hospitals in Latin countries.
Our findings suggest that the mediation role of organizational justice in the relationship between a presenteeism climate and WFC is less salient in the less restrained Latin countries (Smith & Bond, 1999). Essentially, the more extensive freedom of speech and the greater personal control typical of these countries reduces the importance of distributive justice, explaining the indirect relationship between a presenteeism climate (co-worker competitiveness, extra-time valuation and difficulty of replacement) and reduced conflicts at home. This evidence opens avenues for future studies considering the possible role of distributive justice in different countries.

Practical contribution
Our findings also have significant practical implications for managers, essentially by introducing a more comprehensive understanding of how health sector employees perceive justice and how this perception mediates the negative impact of a presenteeism climate on WFC. This may be less of a problem in the Latin countries. Even here, however, HRM policies and practices for improving distributive justice (Greenberg, 2004)  Moreover, companies should consider formal family-friendly programmes to help employees reduce their WFC (Fiksenbaum, 2014). Examples of such programmes include flexible hours, parttime work, job sharing, on-site childcare, flextime, compressed work week, paid maternity/ paternity leave, career breaks, use of employee sick days to attend to family commitments, school holiday and after school care. These family-friendly programmes bring important benefit that makes the organization more competitive for attracting and retaining engaged and high performer workers.

Limitations and suggestions for future research
Our research is cross-sectional in design and the use of self-reported data may raise concerns about method common variance. Specifically, a presenteeism climate appears to be an independent variable that affects WFC, but it is indeed possible that a presenteeism climate may be a cause rather than an effect. However, we carried out measurement and structural invariance tests to ensure validity across the studied samples.
Other methodological issues impact the strengths and limitations of this research. The number of valid responses is small for each subsample. This implies that the statistical power of the estimations is relatively low. Statistically non-significant relationships may be caused by the small number of observations and the heterogeneity of the effects in the aggregate sample.
Moreover, our study population consisted exclusively of employees in the health sector, though it was nationally diverse. This limits the generalization of our findings, although the health sector has a broad range of different occupations. We also found a lack of compatibility in terms of sample characteristics. For example, we only included a few countries in each cluster (Latin vs. Russia and Lebanon countries) due to the convenience of the data collection. Thus, generalizations to other countries should be made cautiously. Moreover, characteristics such as personality traits may have a substantial influence on all variables that are used in the analysis. This limits the conclusions that can be drawn from the estimates.
Additionally, the motives and consequences of presenteeism are most likely heterogeneous across organizations and individuals. In sectors with lower percentages of females the correlations may be different for women vs. men, because men value work more (Simpson, 1998). In the current study, the relatively small sample size limits the possibilities to estimate the relationships for specific subsamples. Considering our evidence, future studies might examine different motives of presenteeism (e.g., chronic vs. non-chronic diseases) or how the type and frequency of illness affects employee performance. Future studies might also consider the relationships between specific groups of workers. Although we use a large sample in our research, greater statistical power for some of the cross-cultural hypothesis testing would be desirable. Future studies could replicate the findings reported here using larger samples of countries and could usefully adopt a wider range of methodologies (both quantitative and qualitative) to fully understand the mechanisms that explain the mediation of organizational justice in the relationship between a presenteeism climate and WFC, including for instance, additional sources of data such as other family members.
Research that considers these constructs will provide a more comprehensive understanding of how a presenteeism climate affects WFC in countries with high and low power distance (Lu et al., 2010). Organizational policies such as flexi-time, lifestyle, demographic and socioeconomic features, as well as other contextual factors, are known to affect WFC Michel et al., 2011;Stock et al., 2016). To gain a more in-depth understanding of the mechanisms through which the work-life balance has an effect on employees' life and organizational climate, future research could also investigate the moderating role of organizational justice in relation to the external context known to influence presenteeism and WFC.

Conclusion
Overall, the results of our study contribute to the field of work-family conflict and presenteeism by showing that organizational justice mediates this relationship differently in different cultures. By showing that we cannot ignore cultural influences on work and family issues, we provide empirical data reinforcing and extending previous theories. In sum, our findings shed new light on the relationship between WFC, a presenteeism climate and distributive justice.