SHIFT WORK AND ABSENTEEISM– THE MEDIATING ROLE OF HEALTH ISSUES, WORK-FAMILY CONFLICT AND ROLE AMBIGUITY

SHIFT WORK AND ABSENTEEISM– THE MEDIATING ROLE OF HEALTH ISSUES, WORK-FAMILY CONFLICT AND ROLE AMBIGUITY

Master thesis, MSc, Human Resource Management University of Groningen, Faculty of Economics and Business

June 29, 2018

P.C.R. VAN ENGELEN Student number: 3534154

Hamburgerstraat 34b 9714JD Groningen

e-mail: p.c.r.van.engelen@student.rug.nl

Supervisor: H.J. van de Brake, MSc

Acknowledgement: I would like to thank Joost van de Brake for his great supervision and

guidance during the writing of my thesis. Furthermore, I would like to thank my colleague

student Liset Mol for the helpful brainstorm sessions and great collaboration regarding both

our theses.

ABSTRACT

In modern society, organizations are expanding their working hours to enable 24-hour operations. To do so, organizations often use shift work, which comes with a demand for flexibility among employees due to irregular working hours. This research looks at three important consequences of shift work on employees; health issues, work-family conflict, and role ambiguity. Building on role theory, it is argued that all three consequences are likely to result into absenteeism, which is considered an important problem in human resource management. This research tests a conceptual model where shift work is positively and indirectly related to absenteeism, mediated by health issues, work-family conflict, and role ambiguity. Furthermore, I will build on autonomy research to propose that the relationships between shift work and both work-family conflict and role ambiguity are moderated by scheduling autonomy. Specifically, I expect that these relationships are less pronounced when scheduling autonomy is higher. Data was collected at a hospital in the South of the

Netherlands, where 367 employees filled in an online questionnaire. The results showed no support for the conceptual model. The research does, however, offer some practical

implications and contributions to literature.

Keywords: shift work, health issues, work-family conflict, role ambiguity, scheduling

autonomy, absenteeism.

INTRODUCTION

Modern society has over recent years changed towards a ’24-hour society’, where time no longer limits business activities (Costa, 2003). Business in all forms is now being done around the clock; not only in services like healthcare, but also in industrial organizations. This demands organizations to expand their working hours, so they can keep up with business. To do so, organizations need to look at a new arrangement of working hours.

The most widely used solution to enable 24-hour activities within organizations is shift work (Costa, 2003). Shift work refers to the arrangement of working hours were employees (individually or in teams) do the same job in succession of each other (Saksvik-Lehouillier, Bjorvatn, Hetland, Sandal, Moen, Magerøy, Harvey, Costa, & Pallesen, 2012). Shift work is related to irregular working hours (Åkerstedt, 1990; Williams, 2008), which demands considerable flexibility from individual employees.

Current literature describes that shift work – and its related need for flexibility – comes with several consequences for shift workers. Specifically, shift work can have both physical and social consequences. First, on a physical level, research has shown a

relationship between shift work and health problems of employees (Åkerstedt, 1990;

Waterhouse, Folkard, & Minors, 1992). Among the physical consequences of shift work, researchers found mental health issues (emotional instability) and adverse cardiovascular effects (e.g., Harrington, 2001). Furthermore, the most repeatedly described physical

consequences of shift work relate to problems with sleeping, such as reduction in quality and quantity of sleep, as well as widespread tiredness (Spurgeon, Harrington & Cooper, 1997).

Second, researchers have investigated the social consequences of shift work. Studies have found that shift workers encounter problems regarding the involvement in social activities (Miles, 1977), as well as keeping in contact with their friends and family (Dunham, 1977).

Thus, current literature has associated shift work with several adverse consequences for

individual employees.

Importantly, however, is that even though studies have reasoned for social consequences of shift work to exist (e.g., Bailyn, Drago & Kochan, 2001; Blachowicz &

Letizia, 2006), these studies have not yet fully succeeded in making clear why these social consequences result from shift work. The present research tries to explain why social consequences emerge from shift work, by attributing them to the different roles that a shift worker has to fulfil. According to role theory (Biddle, 1986), a person has different roles to perform in his or her life, dependent on the situation and setting that this person finds

themselves in. Shift workers have their role at work, next to their roles in their social life with family and friends. Roles differ in responsibilities, expectations, and behaviour, and therefore having different roles can lead to role strain (Jackson & Schuler, 1985). Role strain in turn is found to have a negative effect on workers, as it often creates stress and frustration (Biddle, 1986).

In addition, we know that some people are more tolerant towards shift work than

others (Saksvik, Bjorvatn, Hetland, Sandal & Pallesen, 2011). Hence, shift work may not

always have the same impact on all individuals, and thus the degree of role strain resulting

from shift work will also be different per person. Prior research suggests that one of the most

important influences on people’s reaction to shift work is job autonomy, which is the extent to

which an employee can decide how, where, and when to do their job (Saragih, 2011). Current

literature regularly views autonomy as an important factor to balance work and personal life

(e.g., Valcour, 2007; Voydanoff, 2004). In organization were shift work is highly prevalent, a

particularly important type of autonomy is scheduling autonomy; the extent to which a person

can determine their own work schedule (Pisljar, van der Lippe & den Dulk, 2011). This

research assumes that high levels of scheduling autonomy enable shift workers to have less

role strain, whereas shift workers with low levels of scheduling autonomy are suggested to

face higher levels of role strain.

By establishing that shift work has important physical and social consequences for employees, another question arises: what is the long-term impact of these consequences?

Coming back to the physical consequences of shift work, the most likely result is that employees get physically ill, for example due to sleep deprivation (Knutsson, 2003). When employees get ill, they will most often stay at home for a period to get better. When this happens, there will be absenteeism. However, with the occurrence of absenteeism, it does not necessarily have to be true that the employee is actually physically ill. While there can be other motives for an employee to call in sick, it is the employee’s choice not to show up at work and specify ‘illness’ as the reason (van der Vlist, 1998). For example, employees can choose to stay at home to deal with personal and family matters, caused by the emergence of role strain. According to Rheiner (1982), role strain can be reduced by retracting from the role, which in the case of shift workers can be, for example, by calling in sick for work. This leads to the suggestion that both physical illness and social consequences of shift work can cause absenteeism.

In this research, I test a conceptual model that consists of two paths (see Figure 1).

The first path assumes that shift work is positively and indirectly related to absenteeism,

mediated by health issues. The second path of the conceptual model assumes that shift work is

positively and indirectly related to absenteeism, mediated by role strain (i.e., work-family

conflict and role ambiguity), and where the relationship between shift work and role strain is

moderated by scheduling autonomy. The unique contribution of this research to current

literature is that I tested both paths at the same time. This is important because both paths are

highly distinguishable, yet only the first path has been researched extensively so far. Laying

both path next to each other enables the identification of the importance of both paths when

considering the issue of absenteeism in organizations. Moreover, my research contributes to

shift work literature by looking at the consequences of shift work from a new angle, namely

social consequences. Also, this research will contribute to the role strain literature by

investigating its role in the existence of absenteeism, which at the same time will contribute to absenteeism literature.

The possible practical contribution of my research is that it provides an insight in the consequences of shift work, which enable organizations to consider taking measures to cope with or even eliminate the negative consequences. Furthermore, knowing that absenteeism is considered an important issue to address in human resource management (Bekker, Rutte &

van Rijswijkc, 2009) and a growing problem in organizations due to for example the great costs it comes with (Kocakulah, Kelley, Mitchell & Ruggieri, 2016), the identification of influential factors that cause absenteeism can be used by organizations as a first step in reducing absenteeism.

--- Figure 1 here ---

THEORETICAL FRAMEWORK Shift work

Shift work is often used by organizations to extent the hours of operations beyond that

of conventional working hours, i.e. the standard ‘nine to five’ operational hours (Åkerstedt,

1990). Shift work refers to the arrangement of working hours were employees (individually or

in teams) work in shifts: periods of time in which different (groups of) workers do the same

jobs in succession of each other (Saksvik-Lehouillier, Bjorvatn, Hetland, Sandal, Moen,

Magerøy, Harvey, Costa, & Pallesen, 2012). Shift work is related to irregular working hours,

as well as working in the evening and night hours (Brown, 1975), where employees need to

adapt to. This demands flexibility from employees regarding their working hours.

According to Williams (2008), the usage and amount of shift work depends on the occupation within firms. For example, doctors, nurses and police officers provide services 24 hours per day and therefore are more commonly associated with shift work than occupations where service is provided only during the day, for example office workers or teachers.

Nevertheless, shift work has become more and more common in a variety of industries (Costa, 2003). To illustrate, shift work today enables round-the-clock operations in the

mechanical industry, social services like transport and telecommunications, the food industry, and banking, among other industries (Wyatt & Marriott, 1953).

There are several suggested benefits of shift work for employees. For example, shift work is often related to greater flexibility and a higher salary due to the irregular working hours (Blachowicz & Letizia, 2006). Some employees enjoy shift work and prefer it over traditional working hours, because of those benefits (Hughes & Stone, 2004). However, especially when working in shifts that are outside the ‘normal’ working hours, important challenges are found to occur. The most important adverse consequences of shift work will be discussed later in this report.

There are several ways to operationalize shift work. For example, in current literature, research has often been conducted on shift work in a way that only two groups are compared:

(1) employees that work in shifts, and (2) employees that do not work in shifts (e.g., Brown,

1975; Saksvik-Lehouillier et al., 2012). This research will look at shift work as shift work

variety, because I believe that looking at shift work as a continuum fits better to my research

than approaching shift work as a dichotomous event. This means that I look at shift work in a

way that individuals can have more or less shift work variety. For example, someone who

works only day shifts, only night shifts, or different shifts in a regular schedule, scores low on

shift work. In contrast, someone who works diverse shifts, different from day to day, can be

said to have high shift work variety. In this research I look at the degree of shift work as a

predictor for the degree of consequences that come with shift work. I thereby suggest that employees who score higher on shift work variety, will also experience stronger consequences of shift work.

Physical consequences of shift work

In current literature, shift work is often negatively linked to health issues among employees, such as problems with sleeping and heart problems (e.g., Åkerstedt & Torsvall, 1981; Waterhouse et al., 1992). Costa (1996) suggests that shift work exposes a person to a high level of stress because he or she has to adjust to the changing working hours as quickly as possible, which in the long run can cause permanent problems with sleep and may result into problems with the nervous system. When considering the degree of shift work variety, a person with a high level of shift work consequentially needs to adjust to new working hours more frequently than a person with a low level of shift work. Therefore, the higher the level of shift work, the more problematic the consequences will be for the concerned employee.

According to Tepas & Carvalhais (1990), shift work has been linked to health issues related to the circadian rhythmicity of the body, where sleep, alertness, performance and long- term health were found to concern the major effects. The results of a study by Alfredsson, Karasek & Theorell (1982) associated shift work with higher risks of getting a heart attack.

Shift workers were also found to more commonly suffer from gastrointestinal disorders than non-shift workers (Scott & LaDou, 1994). In addition, Harrington (2001) found that shift work also increases the chance of an employee facing anxiety and depression. Furthermore, another well-recognized possible consequence of shift work is fatigue, which refers to a state of constant tiredness (Pigeon, Sateia, & Ferguson, 2003).

Research has proposed that disruption of circadian rhythms is a potential main reason

for health issues to occur to shift workers (Bøggild & Knutsson, 1999), but also the adoption

or worsening of unhealthy behaviour by shift workers has come forward as a potential factor

leading to health issues among those employees (Shields, 2002). Overall, the most recurring and considered most important physical consequences of shift work relate to problems with sleeping and the effects that these problems have on an employee’s health (e.g., Drake, Roehrs, Richardson, Walsh & Roth, 2004). It is suggested that the effects of shift work are often short term and relate to specific phases of the work schedule. Severe sleeping problems, for example, more often occur in combination with evening or night shifts (Knutsson, 2003).

This suggests that employees that work unregular working hours, experience a more disturbed sleeping rhythm than employees that work regular working hours, due to the need for

adjustment to the varying work schedule. This illustrates that employees with a high level of shift work, are likely to experience more problematic consequences than employees with a low level of shift work.

According to the discussed literature, the degree of shift work thus has a significant impact on an employee’s health conditions. It is suggested that the higher the degree of shift work, the higher the degree of health issues. This leads to the following hypothesis:

Hypothesis 1. Shift work is positively related to health issues.

Absenteeism. When employees encounter health issues and get ill, they will most

likely call in sick for work, which results into absenteeism. That is a problem because

absenteeism can be seen as a person’s non-performance, while the success of organizations is often argued to be highly dependent on the performance of its employees (e.g., Michie &

West, 2004). In literature, there is no universal definition for absenteeism. Harrison and Price

(2003) refer to absenteeism as “the lack of physical presence at a behaviour setting when and

where one is expected to be”. Van der Vlist (1998) proposed a more specified definition for

absenteeism, where he describes it as “the term used when an employee does not show at

work and specifies ‘illness’ as the reason for this absence”. For the purpose of this research,

absenteeism is defined as: the unscheduled absence at work of an employee, where the cause of absence is being attributed to illness.

According to Knutsson (2003), health issues that result from shift work are likely to disappear when the employee reduces the degree of shift work for a certain period of time, for example by taking a day off or taking a longer vacation. Employees that suffer from health problems – regardless of whether these health issues are related to shift work or not – will often call in sick and thus not show up at work until they get better (Harrison & Martocchio, 1998; Halbesleben, Whitman & Crawford, 2014). This will in turn result in the occurrence of absenteeism, leading to the following hypothesis:

Hypothesis 2. Health issues are positively related to absenteeism.

In summary, this research suggests that shift work is positively related to health issues (Hypothesis 1), whereas health issues in turn are positively related to absenteeism (Hypothesis 2). Together, these two predictions suggest that shift work is positively and indirectly related to absenteeism, through health issues. Thereof the following hypothesis derives:

Hypothesis 3. Shift work is positively and indirectly related to absenteeism, mediated by health issues.

Social consequences of shift work

Current literature has suggested other consequences of shift work to be of a social

nature, where shift work is linked to difficulties in employees’ social lives. For example, it is

suggested that shift workers can experience considerable disruption of their family life

(Bailyn, Drago & Kochan, 2001). Furthermore, Williams (2008) explains that shift workers

generally have more difficulty reducing work-life conflicts than non-shift workers. Often,

these difficulties are found to be a result of the irregular working hours that shift workers

operate in; these can make it hard for shift workers to plan social events or to manage family

responsibilities (Levin-Epstein, 2006). At the same time, working in shifts may cause that

shift workers do not have the possibility to frequently communicate with co-workers that work in other shifts, which can create coordination problems among different shifts.

Furthermore, due to this lack of communication, employees can feel less included in the organization as a whole.

So far, studies have discussed social consequences of shift work, but to a limited extent. Most studies did not look at the effects of these social consequences nor at why they result from shift work. This research aims to do so, starting with elaborating on why social consequences result from shift work. Specifically, this research builds on role theory, one of the most important perspectives on the social aspects of work-related strain (Biddle, 1986) to explain the nature of the social consequences. Therefore, in alignment with role theory, this research looks at social consequences of shift work in a way that attributes them to the different roles that an employee has to fulfil. Roles are a person’s social identities that are connected to a set of rights, duties, expectations, norms and behaviours in specific situations (Stryker & Statham, 1985). Since the characteristics of a person’s role depend on the situation that they find themselves in, a person has to perform different roles in their life, sometimes at the same time (Wehner & Thies, 2014). For example, shift workers have to perform their role of employees at work, but they also have their role at home, with their family. These roles differ in responsibilities, expectations and behaviours.

When applying role theory, there are two significant factors that define specific roles.

First, there are role expectations, which are the anticipations of other people about the individual’s behaviour (Rheiner, 1982). Role expectations are, for example, the anticipations of co-workers and family of the shift worker on his or her role. Second, there is role

conception, which refers to the individual’s own perception of their role (Broderick, 1998), thus the perception of the shift worker on their own role. According to role theory,

incompatibility within role expectations and role conception can lead to role strain. There are

two different forms of role strain that can occur for shift workers; role conflict and role ambiguity.

Role conflict happens when a person’s different roles have contradictory demands and role expectations (Miles, 1977). In the case of shift workers, role conflict can occur in the form of work-family conflict, where work related responsibilities interfere with family life (Frone, Russell & Cooper, 1992). Work-family conflict has been a subject of research for many years (E.g., Greenhaus & Parasuraman, 1987; Bedeian, Burke, & Moffett, 1988;

Carlson, Kacmar & William, 2000), and most of these studies have found evidence that suggest that tensions between work and family can lead to a decrease in the psychological well-being of employees (Thomas & Ganster, 1995).

Shift workers can experience difficulties keeping in contact with family or friends due to their irregular working hours. For example, shift workers that work evening or night shifts may experience this difficulty since they have a different life schedule than their friends and family, which makes it harder to find the time to meet up with them (Blachowicz & Letizia, 2006). For employees with a low level of shift work it is easier to combine work and family life since they have a more predictable working schedule. In contrast, employees with a higher level of shift work will experience more work-family conflict, due to the increasing irregularity of their working hours and thus unpredictability of their schedule. Reduced

contact with family and friends can result in feelings of isolation and exclusion (Byron, 2005).

Next to this, role strain is likely to occur due to role conflict since role conflict can create feelings of inability and frustration.

The second form of role strain, role ambiguity, occurs when an individual’s role expectations are vague, and the overview of the individual’s responsibilities is unclear

(Jackson & Schuler, 1985). Research has indicated that uncertainty is an important factor that

can lead to anxiety (Martens, Vealey & Burton, 1990). Even though uncertainty and

ambiguity are not synonymous factors, they are closely linked when concerning roles (Beehr

& Bhagat, 1985). To illustrate, research has suggested that uncertainty follows when employees experience role ambiguity (Beauchamp, Bray, Eys & Carron, 2003). Role strain will then be experienced by the shift workers due to uncertainty, which makes them doubt the value of their work efforts (Beehr, Walsh & Taber, 1976). Role ambiguity has been linked with negative consequences for employees, such as a decrease in confidence about their efficacy, greater feelings of futility, and higher tension (Kahn, Wolfe, Quinn, Snoek &

Rosenthal, 1964; Bandura, 1997). Furthermore, research has found a significant negative correlation between role ambiguity and feelings of personal accomplishment among employees (Acker, 2003).

Shift workers can experience role ambiguity because they are dependent on their colleagues in a sense that they have to pick up the work from the people that worked the previous shift. Due to the use of shifts, it is possible that an individual does not have much contact with colleagues because they work different shifts. Furthermore, shift workers generally do not have one permanent team to work in, but often work with a lot of different co-workers among different shifts. As a result, shift workers can experience a lack of clarity and increased uncertainty since they do not know exactly what co-workers expect from them.

Consequentially, because coordination among colleagues can be difficult due to the lack of

contact, role ambiguity occurs (Smith, Fuqua, Choi & Newman, 2011). However, this

outcome can also depend on the degree of shift work that the employee has. Employees that

have a lower level of shift work often work in similar shifts and will experience less role

ambiguity due to familiarity with the shifts. In contrast, employees with a higher level of shift

work often work in a high variety of shifts and will have less familiarity with those shifts,

resulting in a higher level of role ambiguity.

As discussed, the two forms of role strain, role conflict and role ambiguity, are suggested to derive from or at least are strengthened by shift work; the higher the degree of shift work, the higher the degree of role strain. This idea leads to the following hypotheses:

Hypothesis 4a. Shift work is positively related to work-family conflict.

Hypothesis 4b. Shift work is positively related to role ambiguity.

Scheduling autonomy. It is well known that different people have different reactions

to complex work situations. For example, we know that people react differently to shift work, due to individual factors such as age, gender, and personality (e.g., Barton, 1994; Saksvik, Bjorvatn, Hetland, Sandal & Pallesen 2011). When people react different to shift work, the way in which role strain affects them is also different, due to the relationship between shift work and role strain.

Besides the just mentioned personal differences, there are other factors that can play a role in the different reactions that people have on shift work. Research demonstrates that one of the most important factors is job autonomy (e.g., Spector, 1986; Tremblay & Genin, 2008), which refers to the extent to which an employee has the freedom to decide when, where, and how to do their job (Saragih, 2011). Several researchers acknowledged the positive

relationship between job autonomy and job performance (e.g., Wang & Netemeyer, 2002).

According to Hackman and Oldham (1975), autonomy is one of the most important job characteristics that determines both job motivation and job performance. Davis and Wilson (2000), add that autonomy is one of the crucial factors in increasing job satisfaction and decreasing job stress.

In current literature, autonomy is regularly viewed as an important factor for balancing work and personal life (e.g., Valcour, 2007; Voydanoff, 2004). In the context of this research, this means that autonomy can be viewed as one way to decrease role strain. In current

literature, a distinction in job autonomy is made by defining time/spatial autonomy, which is

the freedom to decide when and where to do one’s job (Pisljar, van der Lippe & den Dulk, 2011). One specific form of job or time/spatial autonomy is particularly relevant in this research: scheduling autonomy (Pisljar, van der Lippe & den Dulk, 2011).

Scheduling autonomy refers to the extent to which a person can determine his or her own work schedule. Thus, it refers to when to do one’s job. When scheduling autonomy is high, it means that employees can plan their own working schedule to a great extent.

Employees that have low scheduling autonomy, by contrast, are more dependent on how others schedule their working hours for them and must adapt to this. This research assumes that employees with high scheduling autonomy will face a lower degree of role strain – both work-family conflict and role ambiguity – than employees with low scheduling autonomy.

First, it is assumed that, when shift work is high, an employee with high scheduling autonomy will face lower work-family conflict. The reason for this assumption is that these employees can schedule working hours according to what is convenient to them and are thus able to adapt their working schedule to their private life responsibilities. Employees that have low scheduling autonomy are, as said, dependent on the schedule that others make for them and they thus have to adapt their private life responsibilities to that schedule. When the level of shift work is high, this asks for a higher degree of flexibility from the employee, which can cause a disruption in their work-life balance (Jackson & Schuler, 1985). Therefore, the

assumption is made that, when shift work is high, an employee with higher scheduling autonomy will face a lower level of work-family conflict, whereas an employee with lower scheduling autonomy will face a higher level of work-family conflict when experiencing the same level of shift work variety.

Second, it is assumed that, when shift work is high, an employee with high scheduling

autonomy will also face lower role ambiguity. As explained earlier, shift workers generally do

not have one permanent team to work in, but rather work with many different people in

different shifts (Smith, Fuqua, Choi & Newman, 2011). Research has shown that when employees have to work with a relatively high number of co-workers, that vary over shifts, they are more likely to experience confusion – or ambiguity – regarding their responsibilities and expectations (Kauppila, 2014). When a shift worker with high shift work variety has high scheduling autonomy, he or she can determine themselves which shifts they want to work in and consequentially with whom they will work. In this way, the employee can choose to pick the shifts where they are familiar with and where their preferred colleagues work in. As a result, the employees are able to build a relationship with those colleagues in which they understand each other’s working values and are more certain about their team roles (Chao, O’Leary, Wolf, Klein & Gardner, 1994). Moreover, it is easier to coordinate tasks because the employees work with the same people and thus know what they expect from them, which eliminates role ambiguity.

In contrast, employees that have low or no scheduling autonomy are not able to

determine which shifts they work in and thus have no influence on with whom they will work.

As a result, employees with low scheduling autonomy are less likely to build good

relationships with their colleagues and have more difficulty knowing what they expect from them. In this case, role ambiguity increases. Therefore, the assumption is made that, when shift work is high, an employee with higher scheduling autonomy will face a lower level of role ambiguity, whereas an employee with lower scheduling autonomy will face a higher level of role ambiguity when experiencing the same level of shift work variety.

The above explained assumptions lead to the idea that the relationship between shift

work and role strain is moderated by the degree of scheduling autonomy that an employee

has. This leads to the following hypotheses:

Hypothesis 5a. Scheduling autonomy moderates the positive relationship between shift work and work-family conflict. This relationship is more pronounced among

employees with lower (rather than higher) scheduling autonomy.

Hypothesis 5b. Scheduling autonomy moderates the positive relationship between shift work and role ambiguity. This relationship is more pronounced among employees with lower (rather than higher) scheduling autonomy.

Absenteeism. As discussed before, health problems among employees can result in

absenteeism. However, there can also be another reason for absenteeism to exist, next to actual illness. Van der Vlist (1998) explains that when there is absenteeism, it is not always clear to what extent there actually is case of illness; when the employee calls in sick, there is absenteeism regardless of the employees’ considerations. So, ‘illness’ refers only to the act of reporting sick, irrespective of its legitimacy and causes (Geurts, 1994). With this in mind, one can to some extent view absenteeism from work as a choice, since it is up to the employees’

interpretation whether they are capable or not to be present at work (Bozionelos & Sun, 2013). Therefore, it is possible that employees report themselves as sick, while they are not actually (physically) ill.

One situation where this can happen, is when an employee does not go to work

because they have to deal with personal and family problems (Cucchiella, Gastaldi & Ranieri,

2014). Social issues can force employees to stay at home for several reasons. For example,

role strain is found to result into stress and frustration (Biddle, 1986), which has a negative

effect on the satisfaction of the employee to go to work. Relatedly, Erickson, Nichols and

Ritter (2000) explain that higher levels of conflict within work-life balance (i.e., role conflict),

have been related to higher self-reported absenteeism. Furthermore, Rheiner (1982) explains

that one way to reduce role strain – both role conflict and role ambiguity – is to retract from

the role relationship, which can for example be by calling in sick. The suggestion can be made

that shift workers, in an attempt to redeem themselves from the consequences of role strain, call in sick for work. Thus, absenteeism can occur regardless of actual illness, in a form of avoidance of work because of role strain. This leads to the following hypotheses:

Hypothesis 6a. Work-family conflict is positively related to absenteeism.

Hypothesis 6b. Role ambiguity is positively related to absenteeism.

Taking it all together, I suggest that shift work is positively related to work-family conflict (Hypothesis 4a), with scheduling autonomy strengthening this relationship

(Hypothesis 5a). Moreover, I suggest a positive relationship between work-family conflict and absenteeism (Hypothesis 6a). Combined, these predictions suggest a positive and indirect relationship between shift work and absenteeism, which is mediated by work-family conflict and moderated by scheduling autonomy. Next to this, I discussed a positive relationship between shift work and role ambiguity (Hypothesis 4b), which is also strengthened by scheduling autonomy (Hypothesis 5b). Combined with the suggested relationship between role ambiguity and absenteeism (Hypothesis 6b), this indicates a positive and indirect relationship between shift work and absenteeism, which is mediated by role ambiguity and moderated by scheduling autonomy. From this, I hypothesize:

Hypothesis 7a. Shift work is positively and indirectly related to absenteeism, mediated by work-family conflict. This indirect relationship is more pronounced among

employees with lower (rather than higher) scheduling autonomy.

Hypothesis 7b. Shift work is positively and indirectly related to absenteeism, mediated by role ambiguity. This indirect relationship is more pronounced among employees with lower (rather than higher) scheduling autonomy.

METHODS

Data collection, design and sample description

I tested my predictions by using multi-source data collected at a hospital in the South of the Netherlands. I approached 777 hospital employees, of which 367 completed the questionnaire (response rate = 47.23%). The employees were contacted via email and were asked to fill in an online, anonymous survey. The survey contained questions and statements about demographical information, health issues, work-family conflict, role ambiguity and scheduling autonomy. Furthermore, I used archival data to capture the participants’ working schedule (i.e., shift work) and absenteeism. The online survey was translated to Dutch by using a double-blind back translation procedure. Data regarding the participants’

demographics were standardized due to the organization’s request to ensure their anonymity.

Measures

Shift work. To measure the degree of shift work variety, archival data consisting the

participants’ working hours in the three months previous to the questionnaire was used to indicate how much percent of their total working time they worked in certain shifts. The shifts were divided over four categories: day shifts, night shifts, overwork (i.e., working more hours than specified in the employee’s contract), and accessibility shifts (i.e., shifts where the employee is not at work, yet has to be available to be called up in case he or she is needed at work). I thereafter used Blau’s (1977) index of heterogeneity to measure the degree of variation in shifts per employee, according to the archival data. The formula for Blau’s index is 1 - ∑ Pi

(Richard, 2000), where P is the proportion of the total hours that the participant worked in those three months, and i is the number of different shifts that he or she has worked in. The index ranges from 0 to 1, where 0 means no variety and 1 means maximum variety.

To illustrate, when a participant has an index of 0, he or she only worked shifts in one of the four categories in those three months. A participant with an index of .50, worked half of their hours in one shift category, for example day shifts, and the other half in another shift

category, for example night shifts.

Health issues. To capture health issues, I used one item in the questionnaire where I

asked: “How often have you suffered from health complaints during your work in the past year? Such as: gastrointestinal problems, restlessness / nervousness, difficulty concentrating, pain in the (lower) back, neck pain, dizziness, exhaustion, headache, or heartburn.” (cf.

Pereira & Elfering, 2014). The options regarding the frequency of occurrence of these health issues were: ‘never’, ‘once per year’, ‘once per half-year’, ‘once per month’, ‘once per week’,

‘once per day’, and ‘multiple times per day’. The item was coded so that a higher frequency indicated a higher degree of health issues.

Work-family conflict. Within the questionnaire, I used 5 items to capture the concept

of work-family conflict (cf. Netemeyer, Boles & McMurrian, 1996). Participants were asked to respond to these items by indicating whether they agreed or disagreed with the statement. A 7-point Likert scale was used, ranging from ‘totally disagree’ to ‘totally agree’. An example of the items asked is ‘the demands of my work disrupt my home- and family life’. The items were coded so that ‘totally disagree’ would mean no work-family conflict and ‘totally agree’

would mean a high level of work-family conflict. The items formed a reliable scale (Cronbach’s alpha > .80).

Role ambiguity. To capture role ambiguity, I used three items consisting statements

about role clarity, where participants were asked to indicate how clear certain aspects of their

job are for them (cf. Jokisaari & Nurmi, 2009). The items were measured through a 7-point

Likert scale, ranging from ‘very unclear’ to ‘very clear’. An example of these items is ‘how

clear are your responsibilities to you?’. Since these items indicate role clarity, rather than role

ambiguity, the answers were adversely coded so that ‘very unclear’ means a higher score on

role ambiguity and ‘very clear’ means a lower score on role ambiguity. Together, these items

formed a reliable scale (Cronbach’s alpha > .80).

Scheduling autonomy. I used archival data to measure the participants’ scheduling

autonomy as a binary variable (0-1). In the hospital, it is known which employees have the opportunity to make their own schedule and which employee do not have this opportunity.

Among the participants, 118 employees were known to have scheduling autonomy and 249 employees were known not to have scheduling autonomy. I coded the data so that a score of 1 means that the employee is able to make his or her own schedule, and thus has scheduling autonomy. A score of 0 indicates no scheduling autonomy.

Absenteeism. To measure the degree of absenteeism, I used archival data on the

reported hours of sick leave per employee. I reported the total amount of hours of sick-leave per participant, over three months.

Control variables

To gain more insight on the background of the participants of the study and ensure that these factors had no influence on the relationships between the variables in my conceptual model, I measured some control variables. The control variables that I included were gender, age, and workload.

Gender. I included the control variable gender because current literature has found

that women have increased difficulties with shift work compared to men, related to their domestic obligations (Spurgeon, Harrington & Cooper, 1997). Furthermore, research has assumed that shift work has more adverse effects on women’s health, especially regarding their periodical hormonal activity (Costa, 1996). Moreover, it was found that female shift workers more often complain about tiredness or sleeping problems (Brown & Charles, 1982).

Gender was captured with the use of archival data. The scores were standardized for

anonymity reasons. Due to the standardization of the scores, I could not trace back which

standardize score indicated male or female. Therefore, I was able to control for this variable,

however I was unable to interpret the outcomes of the analyses, regarding gender.

Age. I included the control variable age because current literature has suggested the

relationship between age and an employee’s tolerance to shiftwork (Harrington, 2001).

Research has shown evidence of older workers tolerating shift work less well than younger colleagues (Torsvall, Åkerstedt & Gillberg, 1981), since the older workers are also more susceptible to sleep disorders (Marquie & Foret, 1999). The variable age was measured in years via a single question within the questionnaire. The scores were standardized for anonymity reasons.

Workload. I included workload as a control variable because previous research

suggests that employees with a higher workload more often display higher levels of absenteeism (Rotchford & Roberts, 1982). I used archival data to capture an employee’s workload by collecting the employee’s overall contractual working time over three months, measured in full-time equivalent (FTE). The scores were standardized for anonymity reasons.

Analysis

The gained data was analysed with the help of the statistical program SPSS. To estimate the direct and indirect effects of the mediation of health issues between shift work and absenteeism, model 4 of the modelling tool PROCESS macro (Hayes, 2009) was used in SPSS. The moderated-mediation effects of work-family conflict and role ambiguity on the relationship between shift work and absenteeism, moderated by scheduling autonomy were estimated by using of model 7 of the modelling tool PROCESS macro (Hayes, 2009) in SPSS.

RESULTS Descriptive statistics

In table 1, the means, standard deviations and bivariate correlations for all study

variables are presented. In contrast with my expectations, I found no significant correlation

between shift work and health issues (r = -.01, p > .05). I did find preliminary support for a

positive correlation between health issues and absenteeism (r = .14, p < .01). This indicates a relationship where a higher level of health issues leads to a higher level of absenteeism.

The expected positive correlation between shift work and work-family conflict was also found to be significant (r = .17, p < .01), indicating that a higher level of shift work variety results into higher levels of work-family conflict. The correlation between work- family conflict and absenteeism, however, was not found to be significant, even though this was expected (r = .08, p > 0.05). The correlation between shift work and role ambiguity was also not found to be significant (r = -.06, p >0.5), and neither was the correlation between role ambiguity and absenteeism (r = .04, p > 0.05).

Interestingly, I found positive correlations between work-family conflict and the other two mediators; health issues (r = .30, p < .01) and role ambiguity (r = .22, p < .01). This suggests that a higher the level of work-family conflict is connected to higher levels of health issues and role ambiguity. In addition, work-family conflict was also found to negatively correlate with the moderator scheduling autonomy (r = -.12, p < .05), indicating that a higher level of scheduling autonomy directly leads to a lower level of work-family conflict.

Importantly, I found a negative correlation between shift work and absenteeism (r = - .11, p < .05), where I expected to find no direct correlation here, yet only an indirect and positive relationship. The correlation indicates that when the level of shift work variety is higher, the level of absenteeism is lower.

Regarding control variables, I expected gender to be correlated to health issues,

however I did not find support for that in my data (r = -.09, p > .05). I did find support for the

expected correlation between gender and absenteeism (r = .12, p < .05). I note, however, that I

cannot draw any conclusions from this correlation, since data regarding the variable gender

was standardized and I thus am unable to interpret the scores for gender. I therefore do not

know whether this correlation suggests a relationship where females score higher on

absenteeism or males. Moreover, I found a negative correlation between age and shift work (r

= -.16, p < .01), indicating that the higher the employee’s age, the lower their shift work variety. Surprisingly, I also did not find a significant correlation between age and health issues (r = -.06, p > .05), where I did expect to find a positive correlation between the two variables. Lastly, I expected to find a positive correlation between workload and absenteeism, which was not supported by my results (r = -.01, p > .05).

--- Table 1 here --- Hypothesis testing

Hypothesis 1 predicted a positive relationship between shift work and health issues, where a higher degree of shift work variety would lead to more health issues. The results in Table 2 show that there was no support found for this positive relationship (B = -.04, p > .05).

Therefore, Hypothesis 1 is not supported.

Hypothesis 2 predicted a positive relationship between health issues an absenteeism, indicating that a higher degree of health issues leads to a higher level of absenteeism. As shown in Table 2, this relationship was found significant (B = .05, p < .05). Hence,

Hypothesis 2 is supported. Furthermore, Hypothesis 3 predicted that health issues mediates the relationship between shift work and absenteeism. As illustrated in Table 2, the mediation analysis showed no support for an indirect relationship between shift work and absenteeism, via health issues (B = -.00, CI = -.02 to .01). Thus, Hypothesis 3 is not supported.

---

Table 2 here

---

Hypothesis 4a predicted the positive relationship between shift work and work-family conflict, where more shift work variety is suggested to lead to a higher degree of work-family conflict. The results in Table 3 show significant support for this prediction (B = .21, p < ,05).

Therefore, Hypothesis 4a is supported. Hypothesis 4b predicted the positive relationship between shift work and role ambiguity, where more shift work variety is suggested to lead to a higher degree of role ambiguity. The results in Table 4 show no significant relationship between these two variables (B = -.06, p > .05). Therefore, Hypothesis 4b is not supported.

Hypothesis 5a predicted the moderation of scheduling autonomy on the relationship between shift work and work-family conflict, where this relationship is more pronounced for employees with a lower level of scheduling autonomy. As shown in Table 3, the interaction term for shift work and scheduling autonomy was not found to be significantly related to work-family conflict (B = .16, p >.05). Thus, Hypothesis 5a is not supported by my results.

Hypothesis 5b predicted the moderation of scheduling autonomy on the relationship between shift work and role ambiguity, where this relationship is more pronounced for employees with a lower level of scheduling autonomy. Table 4 shows that the interaction term for shift work and scheduling autonomy was not found to be significantly related to role ambiguity (B = - .11, p >.05). Thus, my results do not support Hypothesis 5b.

Hypothesis 6a predicted the positive relationship between work-family conflict and

absenteeism, meaning that a higher degree of work-family conflict results in a higher level of

absenteeism. The results in Table 3 show no significant relationship between these two

variables (B = .03, p > .05). Hence, Hypothesis 6a is not supported. Hypothesis 6b predicted

the positive relationship between role ambiguity and absenteeism, meaning that more role

ambiguity results in a higher level of absenteeism. The results in Table 4 show no support for

this prediction (B = -.03, p > .05). Hence, Hypothesis 6b is not supported.

Hypothesis 7a predicted the mediation of work-family conflict in the relationship between shift work and absenteeism, where the relationship between shift work and work- family conflict is moderated by scheduling autonomy. The moderated-mediation analysis in Table 3 shows no significant moderated-mediation relationship between shift work and absenteeism, via work-family conflict and moderated by scheduling autonomy (index = .00, CI = -.00 to .03). Therefore, Hypothesis 7a is not supported by my results. Hypothesis 7b predicted the mediation of role ambiguity in the relationship between shift work and

absenteeism, where the relationship between shift work and role ambiguity is moderated by scheduling autonomy. The results in Table 4 show no significant moderated-mediation relationship between shift work and absenteeism, via role ambiguity and moderated by scheduling autonomy (index = .00, CI = -.00 to .01). Thus, my results do not support Hypothesis 7b.

--- Table 3 & 4 here --- Additional analysis

With regard to absenteeism, exploratory analyses indicated that this variable followed

a non-normal distribution (see Figure 2). With this distribution, common assumptions of OLS

analysis are violated (Cameron & Trivedi, 1998). To ensure that the results reported above

were valid, I re-examined the relationship between shift work and absenteeism, mediated by

health issues, work-family conflict, and role ambiguity, with a recoded variable and a more

appropriate statistical method. Specifically, I recoded the variable absenteeism into a

dichotomous variable where one can either score 0 or 1, where 0 means that there was no

absenteeism recorded and 1 means that there was absenteeism recorded. After transforming

the variable, I used logistic regression in SPSS to predict absenteeism, using the same

variables as reported above. However, estimating absenteeism using this new variable in my model resulted in similar parameter estimates and significance levels as the previous analysis.

--- Figure 2 here --- DISCUSSION Summary

Current literature has linked shift work to several consequences for employees, such as health issues (e.g., Costa, 1996; Harrington, 2001; Knutsson, 2003) and role strain – namely work-family conflict (e.g., Blachowicz & Letizia, 2006; Williams, 2008) and role ambiguity (e.g., Jackson & Schuler, 1985; Smith, Fuqua, Choi & Newman, 2011). Literature also indicates that the relationship between shift work and the social consequences role strain is influenced by the level of scheduling autonomy that the employee has, where a higher level of scheduling autonomy would lead to a less pronounced relationship between shift work and role strain (e.g., Valcour, 2007; Vaydanoff, 2004). In turn, the consequences of shift work are suggested to lead to absenteeism (e.g., Rheiner, 1982; Harrison & Martocchio, 1998;

Erickson, Nichols & Ritter, 2000; Halbesleben, Whitman & Crawford, 2014).

I tested my predictions in a conceptual model, illustrated in Figure 1. The findings of my research show no support for this model. However, they do show support for the

relationship between health issues and absenteeism, and the relationship between shift work and work-family conflict. Thus, in line with previous research, the present study shows that a higher level of health issues leads to more absenteeism, and a higher level of shift work variety leads to a higher level of work-family conflict.

Theoretical implications

In my research I found that shift work variety is positively related to work-family conflict, meaning that more shift work variety leads to a higher level of work-family conflict.

This finding has two important contributions to literature.

First, it contributes to work-family conflict literature since it discovered a rather new cause for work-family conflict. In current literature, researchers have regularly found that work-family conflict is caused by a disbalance between work and family life (e.g., Dunham, 1977; Frone, Russell & Cooper, 1992; Carlson, Kacmar & William, 2000). The direct link between work-family conflict and shift work has been discussed (Bailyn, Drago & Kochan, 2001), yet not very often. The present study found that employees with a higher degree of shift work variety show higher levels of work-family conflict than employees with lower shift work variety. With that finding, the present study goes further into depth regarding the

relationship between shift work and work-family conflict, by discovering that not just shift work as a dichotomous factor, but rather the degree of shift work variety plays a role in the occurrence of work-family conflict.

Second, as the present research found a relationship between shift work and work- family conflict, it indicates that work-family conflict is an issue that derives from shift work.

Besides contributing to work-family conflict literature, this finding relatedly contributes to the shift work literature, because I looked at the consequences of shift work from a different angle, namely from a social perspective, whereas current literature focusses mainly on the physical consequences of shift work (e.g., Åkerstedt, 1990; Spurgeon, Harrington & Cooper, 1997; Harrington, 2001; Drake, et al., 2004). Social consequences are important consequences of shift work, which have slightly been discussed and argued for in literature before (e.g., Dunham, 1977), yet the present study has found evidence to support these arguments.

Moreover, the present study brings another contribution regarding shift work

literature. In my research, I measure shift work in a setting where everyone works in shifts,

yet some with more variety than others. I use a different focus on shift work, namely one that looks at shift work as a continuum instead of a dichotomous event. That means that within the present study, participants can score higher or lower on shift work. Previous studies have looked at shift work as a dichotomous variable, meaning that the participants are divided in two categories: (1) employees that work in shifts, and (2) employees that do not work in shifts (e.g., Costa, 2003; Williams, 2008; Saksvik-Lehouillier et al., 2012). My approach to shift work offers a new perspective and therefore a contribution to shift work literature.

Another finding of my research is that health issues are positively related to objective absenteeism registrations. With that finding, the present study further strengthens the linkage between health issues and absenteeism, as has been described frequently in current literature (e.g.: Harrison & Martocchio, 1998; Halbesleben, Whitman & Crawford, 2014). My research confirms the well-research relationship, where more health issues leads to a higher level of absenteeism. In addition, I used archival data to estimate the participants’ absenteeism, which is objective data and therefore could be a more trustworthy approach than asking the

participants about their absenteeism in a questionnaire.

Limitations and future research

Despite the usage of multi-source data, which strengthens my methodology, my research also holds a number of potential limitations. To start, I tested my hypotheses in a sample of employees within one organization in one country, the Netherlands. This limits the generalizability of my results since it only covers one specific organizational and cultural context. It is possible that other types of organizations would show different results. For example, sample organization of the research is a hospital, which generally is a highly complex organization with lots of dynamics. This can already have an influence on the level of work-family conflict, role ambiguity, or health issues that an employee experiences.

Employees in organizations with less complexity or dynamics might experience less of these

issues, regardless of the level of shift work variety they have. Future research could enable generalization when constructively replicating the present study in other contexts and comparing the results.

Another limitation of my research can be the time frame in which the research was conducted. Due to the relatively short research period, it was not possible to capture a change of shift work variety per employee over time. It can for example be interesting to see if one participant that score high on work-family conflict, will in fact experience this consequence less when his or her shift work variety decreases. Also, it can be interesting to see if a participant who scores high on shift work variety will show less health issues or role strain after one or two years with the same high level of shift work variety, possibly because he or she simply got used to it and thus his or her tolerance to shift work increased over time (e.g., Folkard & Hunt, 2000). Therefore, a future research opportunity is to use a longitudinal study to estimate the effect of change in shift work variety and the effect of having the same level of shift work variety over time.

Future research can also be dedicated to the relationship between shift work and health issues. As discussed, previous literature has suggested that shift work leads to health issues among employees (e.g., Åkerstedt & Torsvall, 1981; Costa, 1996; Harrington, 2001;

Knutsson, 2003), however, this is not supported by my results. Potentially, this can be caused by the way I looked at shift work, namely as a continuum measuring shift work variety. It is possible that the degree of shift work variation has no influence on an employee’s health, but rather the fact that this employee works in shifts – regardless of the variety. Future studies could therefore use the variable shift work in the traditional way to discover whether this will show different results and confirm the relationship that theory suggests. Relatedly, the

mediation of health issues in the relationship between shift work and absenteeism was also

not found and can in this way perhaps be discovered in future research as well.

Furthermore, this research suggested a relationship between shift work and role ambiguity, where a higher degree of shift work variety was suggested to lead to a higher degree of role ambiguity. Despite that this suggestion builds on existing literature (e.g., Jackson & Schuler, 1985; Smith, Fuqua, Choi & Newman, 2011), the relationship was not found in my results. Consequently, due to the missing relationship between shift work and role ambiguity, the moderating effect on scheduling autonomy is also not found. A potential reason for the missing relationship between shift work and role ambiguity in my results could be the existence of organizational procedures. Due to the nature of the sample organization – a hospital – it is reasonable to assume that employees have to follow strict guidelines,

considering the need for specific procedures within healthcare (e.g. regarding hygiene) The existence of these guidelines could eliminate the presence of role ambiguity, since the work needs to be done in the exact same way, regardless of the shifts or colleagues that one works with. Further research within different organizational contexts could reveal whether this is the case or not and perhaps then discover a relationship between shift work and role ambiguity.

Interestingly, the moderated effect of scheduling autonomy on the relationship between shift work and work-family conflict has not been found in my research as theory would suggest (e.g., Valcour, 2007; Vaydanoff, 2004), however, I did find a direct

relationship between scheduling autonomy and work-family conflict. This relationship was

negative, indicating that a higher level of scheduling autonomy directly leads to a lower level

of work-family conflict. This finding gives reason to believe that scheduling autonomy thus

has an influence on work-family conflict, yet regardless of whether the employee works in a

high or low variety in shifts (or not in shifts at all). Scheduling autonomy is found to be an

important factor in balancing work and personal life (e.g., Pearson & Moomaw, 2005; Pisljar,

van der Lippe & den Dulk, 2011) and according to my results, shift work – or the variety in

shifts – has nothing to do with this relationship. The relationship, however is interesting to further research, since it can give implications on how to deal with work-family conflict.

Another opportunity for future research is to look at the consequences of work-family conflict and role ambiguity. Literature suggested that a higher level of role strain leads to a higher level of absenteeism (e.g., Rheiner, 1982; Erickson, Nichols & Ritter, 2000), however this is not supported by my results, for either two forms of role strain. Relatedly, the

mediation of work-family conflict and role ambiguity between shift work and absenteeism, moderated by scheduling autonomy, is also not found in my research. A potential reason why I did not find a relationship between role strain and absenteeism, is that role strain leads to absenteeism indirectly, via other factors. For example, I found a positive correlation between work-family conflict and health issues, indicating that a higher level of work-family conflict leads to a higher level of health issues. Some scholars also describe the linkage between work-family conflict and health issues, for example due to the physical consequences of stress (e.g., DeLongis, Folkman & Lazarus, 1988). Investigating this matter further indicates a conceptual model where work-family conflict leads to absenteeism, yet mediated by health issues, which can be interesting for future studies on these topics.

Practical implications

My research offers some implications for practitioners. First, I found that variation in

shift work is positively related to work-family conflict. This suggests that organizations that

aim to support their employees in balancing their work and family life, can therefore take into

account shift work variety. They can do this for example when making work schedules for

their employees – assuming that the employees have no scheduling autonomy – and by trying

to keep the level of shift work variety lower rather than higher. Smaller organizations can also

consider investing time in figuring out the best level of shift work variety per employee, by

consulting with them to discover their preferences.

Furthermore, my research offers implications for practitioners that aim to get a grip on absenteeism. Taking into account that my findings confirmed health issues to be a predictor for absenteeism, organizations can consider to more frequently make inquiries about the health status of their employees. In this way, they can create a better view on the course of events regarding health, and consequently create a better view on when to expect sick leave.

Thus, my findings do no give a solution to eliminate absenteeism, but rather a way to get a grip on the issue.

CONCLUSION

This research tested a conceptual model where shift work is positively and indirectly related to absenteeism, mediated by health issues, work-family conflict, and role ambiguity.

Moreover, I built on autonomy research to propose that the relationships between shift work

and the two forms of role strain are moderated by scheduling autonomy. Even though the

results of my research showed no support for the conceptual model, my research did provide

some interesting practical implications regarding shift work, work-family conflict and

absenteeism. I hope that these implications will help practitioners in the future. Furthermore,

the present research has opened several doors for interesting future research on a variety of

topics.

REFERENCES

Acker, G. M. (2003). Role conflict and ambiguity: Do they predict burnout among mental health service providers? Social Work in Mental Health, 1(3), 63-80.

Åkerstedt, T. (1990). Psychological and psychophysiological effects of shift work. Scandinavian Journal of Work, Environment & Health, 67-73.

Åkerstedt, T., Torsvall, L. (1981). Shift work. Shift- dependent well-being and individual differences. Ergonomics, 24, 265–73.

Alfredsson, L., Karasek, R., & Theorell, T. (1982). Myocardial infarction risk and psychosocial work environment: an analysis of the male Swedish working force. Social science & medicine, 16(4), 463-467.

Bailyn, L., Drago, R., & Kochan, T. A. (2001). Integrating work and family life. A Holistic Approach, A Report of the Sloan Work-Family Policy Network: MIT, Sloan School of Management.

Bandura, A. (1997). Self-efficacy: the exercise of control. New York: Freeman.

Barton, J. (1994). Choosing to work at night: a moderating influence on individual tolerance to shift work. Journal of Applied Psychology, 79(3), 449-454.

Beauchamp, M. R., Bray, S. R., Eys, M. A., & Carron, A. V. (2003). The effect of role ambiguity on competitive state anxiety. Journal of Sport and Exercise Psychology, 25(1), 77-92.

Bedeian, A. G., Burke, B. G., & Moffett, R. G. (1988). Outcomes of work-family conflict

among married male and female professionals. Journal of Management, 14, 475-491.

Beehr, T. A., & Bhagat, R. S. (1985). Introduction to human stress and cognition in organizations. Human stress and cognition in organizations, 3, 19.

Beehr, T. A., Walsh, J. T., & Taber, T. D. (1976). Relationships of stress to individually and organizationally valued states: Higher order needs as a moderator. Journal of applied psychology, 61(1), 41.

Bekker, M. H., Rutte, C. G., & Van Rijswijk, K. (2009). Sickness absence: A gender-focused review. Psychology, health & medicine, 14(4), 405-418.

Biddle, B.J. (1986). Recent development in role theory. Annual Review of Sociology, 12, 67- 92.

Blachowicz, E., & Letizia, M. (2006). The challenges of shift work. Medsurg Nursing, 15(5), 274.

Blau, P. M. (1977). Inequality and heterogeneity: A primitive theory of social structure (Vol.

7). New York: Free Press.

Bøggild, H., & Knutsson, A. (1999). Shift work, risk factors and cardiovascular disease.

Scandinavian journal of work, environment & health, 85-99.

Bozionelos N., & Sun J. M. 2013. Do labour markets impact whether sick employees call in sick? Academy of Management Perspectives, 27(3).

Broderick, A. J. (1998). Role theory, role management and service performance. Journal of services marketing, 12(5), 348-361.

Brown, D. (1975). Shiftwork: a survey of the sociological implications of studies of male

shiftworkers. Journal of Occupational and Organizational Psychology, 48(4), 231-

240.