Fleur Klijn Bachelor Thesis

37  Download (0)

Full text


Do dominant leaders cause work stress? A study on this relationship within the remote work environment.

Bachelor Thesis Fleur Klijn

Student number: 12412155

Supervisor: Dr. D. M. Dekker

Bachelor thesis Business Administration

University of Amsterdam

June 30, 2021

Words: 6665


Statement of originality

This document is written by Fleur Klijn who declares to take full responsibility for the contents of this document.

I declare that the text and the work presented in this document are original and that no sources other than those mentioned in the text and its references have been used in creating it.

The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.



The global pandemic caused a lot of changes and challenges for both followers and their leaders. One of the biggest changes was working remotely. Previous research discovered that leadership styles affect how people feel at work. There has been no research on dominant leadership on followers’ work stress during a pandemic. Quantitative research was conducted to investigate if starting a job during the pandemic affects the relationship between dominant leadership and work stress for followers working remotely. This study proposes that a more dominant leader increases followers' work stress. Moreover, it suggests that the starting point of a followers' job influences this relationship. More specifically, it hypothesizes that there is a positive relationship between dominant leadership and work stress within the remote working environment (hypothesis 1) and that this relationship is stronger if a follower started their job during the pandemic (hypothesis 2). The hypotheses were tested with a sample of 88 employees who filled in an online survey. Hypothesis 1 was supported, and hypothesis 2 was not supported. This suggests that dominant leaders cause more work stress. However, starting your job before or during the pandemic did not influence the relationship between dominant leadership and work stress.


Table of contents

Statement of originality 2

Abstract 3

1. Introduction 5

2. Theoretical Framework 6

2.1. Corona and working remotely 6

2.2. Follower work stress 7

2.3. Dominant leadership 8

2.4. Starting a new job 9

3. Research Method 11

3.1. Design 11

3.2. Sampling and participants 11

3.3. Procedure 12

3.4. Measurements 12

3.4.1. Main variables 12

3.4.2. Control variables 13

3.5. Analytical plan 14

4. Results 14

4.1. Correlations 14

4.2. Assumptions testing 15

4.3. Hypotheses testing 16

5. Discussion 17

5.1. Summary findings and meaning 17

5.2. Other findings and alternative explanations 18

5.3. Limitations and future research 19

5.4. Practical implications 20

5.5. Conclusion 21

Reference list 22

Appendices 27


1. Introduction

In February 2020, life in the Netherlands changed. For months there was no fear until the virus came closer and it struck our country. Now, leaving the third wave, COVID-19 has been dominating our daily lives for more than a year. Mouth masks, disinfection gel and one- and-a-half-meter society have become part of our new normal. The pandemic changed many aspects of our life. Regarding the work side, it has caused many challenges and difficulties for both supervisors and followers. Working remotely, which has been happening for many years, is now obligatory as it is prohibited to work with as many people in one office as there were. The remote and virtual work environment can cause stress for followers as they have never been in this situation. For leaders, it is therefore important to look at their leadership styles to know how followers are affected by it in a remote environment. Specifically, in this paper, how it affects followers’ work-related stress.

There has been much research on followers’ work stress and what the causes and effects of work stress are (Stranks, 2015; Hon & Kim, 2007; Kahn & Byosiere, 1992). Work stress causes can be categorized in three main aspects which are the physical environment, the organization and by the way the organization is led. The effects of work stress are lower job performance, poorer health, and lower physiological well-being (Kahn & Byosiere, 1992;

Terry, Nielsen & Perchard, 1993; Vrijkotte, van Doornen, de Geus, 2000). These are all components that lower the overall performance of a company. Therefore, it is important to address work stress within a company.

Leadership styles also influence work stress of a follower. For example, a considerate leader results in lower stress of followers (Lopez, Green, Carmody-Bubb, Kodatt, 2010) Furthermore, Bhatti, Shar, Shaikh and Nazar (2010) found that a leadership style which includes misbehaving results in more stress at work for followers. However, there is not much information on a dominant leadership style in relation to followers’ work stress, and as research has shown, it is likely that there will be more dominant leaders in time of uncertainty, such as in the current pandemic (Kakkar & Sivanathan, 2017). Therefore, it is important to know how this leadership style affects followers’ work stress. Furthermore, as working remotely will be integrated in our lives for a longer time, it is important to understand how specific concepts are influenced by this.

Additionally, it is assumed that starting a new job during the pandemic can strengthen the relationship between dominant leadership and followers’ work stress. As both trust and strong relationship building are indicators of followers’ work stress, and building those is more complex in a remote work environment, new followers, who started their job working remotely,


will experience more work stress (Avolio, 1999; Jermier & Kerr, 1997; Nelson, Basu & Purdie, 1998; Davis & Schoorman, 1995; Staples, 2001). Besides, building a relationship and trust remotely with a dominant leader also showed to be more challenging (Czech & Forward, 2010).

Therefore, as little research has been done on dominant leadership and work stress of followers, who also start their job during the pandemic, within the remote work environment, the following research question has been formulated: Does dominant leadership positively relate to work stress of a follower within the remote work environment? And is this effect stronger when a follower started its job during the corona crisis?

This study aims to add knowledge to the existing literature of leadership styles in relation to work stress of a follower. This thesis contributes to science as it looks at already consisting and investigated concepts, however now within a corona and remote environment.

Little to no research has been done on this pandemic on many concepts. Therefore, it adds to the current knowledge. Furthermore, the study gives insights to supervisors and leaders who are currently battling working remotely. It gives them an understanding of how a specific leadership style influences the work stress of their employees. It could make them rethink and alter their leadership style to create an efficient and safe environment for employees to work in.

It also gives them information on whether or not to put more effort into employees who started their job during the pandemic if it shows that they get even more stressed by a dominant leader.

The following section, section two, will include the theoretical framework, which discusses all the relevant and existing literature about the concepts discussed in this thesis. It will also introduce the hypotheses. Furthermore, in section three, the research method will be discussed. After analysing the data, the results part follows in section four. In the last section, section five, there will be a discussion of the results, including suggestions for future research and the limitations of this research. It finishes with an answer to the research question.

2. Theoretical Framework

This section discusses the relevant literature related to remote work, dominant leadership, work stress and what the influence can be of starting your job before or during the pandemic. The hypotheses are formulated, and these are supported by a conceptual model at the end.

2.1 Corona & Working Remotely

A crisis situation is dominating our daily lives right now: The COVID-19 pandemic.

Working remotely and virtual is the new ‘normal’, which resulted in more remote leaders.

Remote leadership is explained as an electronically supported communication between


physically isolated leaders and followers (Kelloway, Barlong, Kelly & Comotis, 2003). The improved availability of the internet makes working remotely a new feasible method of work (Wang, Ang, Jiang & Wu, 2021). This new work environment is a situation where there is physical distance and where face-to-face communication is almost non-existent because they are in different locations (Zimmermann, 2008).

According to Mann and Holdsworth (2003), the greatest advantage of working remotely is that there is increased flexibility and increased productivity. However, there are also multiple disadvantages. The first disadvantage and challenge of remote work is that making relationships with their followers becomes more difficult for leaders (Avolio, 1999; Kerr & Jermier, 1978).

It has also been found that physical distance has a negative effect on performance (Howell and Hall-Merenda, 1999) besides having a negative impact on leadership. Wang et al. (2021) found the same result and explained that the absence of social interaction together with a lower quality of communication could result in employees that are not able to adapt to the new situation.

When employees cannot adjust to the new situation, job performance will decrease. Other challenges of remote work include collaborating and communicating with others, and finally it is found difficult to separate home and work life (Flores, 2019).

As we are now in the situation that working remotely is almost obligatory, it is important to look at its effects on employees. As Vinkers et al. (2020) stated that stress is a normal reaction to the pandemic, therefore it is even more interesting to look at a more specific type of stress within the remote work environment, namely work-related stress.

2.2 Follower work stress

The Health and Safety Executive describes that employees experience work stress when they cannot cope with the demands and pressures put on them within their job (Stranks, 2015).

He also described that stress is related to how people adapt to changes in their life. New jobs, promotion of suddenly working from home are all examples of this. If a person is not able to cope with these changes, more stress is expected. Stress, in general, is not always bad, as some people need a certain amount of positive stress to carry out their tasks. An example is challenge- related stress, discussed by Hon & Kim, 2007). This is often defined as ‘good stress’, resulting in a feeling of achievement and fulfilment. However, this kind of stress can also result in work overload and time pressures (Hon & Kim, 2007).

There has also been research on hindrance-related stress, which is bad stress (Podsakoff, Lepine & Lepine, 2007). This kind of stress is related to the interference of the ability of a person to achieve specific goals. Hindrance-related stressors include role conflict, concerns


about job security and role ambiguity. Many different kinds of stress have been researched.

However, the focus in this study is on work stress only.

Multiple causes for work stress have been found. There are various dimensions from which stress could be caused. Stranks (2015) divided the causes into three categories. Stress caused by the physical environment, the organisation or by the way the organisation is led. He stated that within the physical environment factors such as privacy and noise levels result in higher stress levels. Within the organisational level, no time to adjust to change, insufficient training, job insecurity and workload are examples of factors influencing work stress. Last, leadership ways will influence work stress as inconsistency of style, emphasising competitiveness and poor information sharing can cause work stress.

Furthermore, work stress also has many effects. A well-known effect is that it is negatively related to job response and organisational effectiveness (Kahn & Byosiere, 1992).

High levels of stress also result in poorer psychological well-being (Terry, Nielsen & Perchard, 1993). The last significant effect is that it negatively affects physical health. Work stress negatively influences the heart rate and blood pressure (Vrijkotte, van Doornen, de Geus, 2000).

As the effects of work stress can be tremendous to an organization, there needs to be a closer look at the causes to tackle work stress. As the way an organisation is led also influences work stress, there will be a closer look in this study at how a specific leadership style influences followers’ work stress.

2. 3 Dominant leadership

A dominant leadership style is identified as a leader who uses power, fear and coercion (De Waal-Andrews, Gregg and Lammers, 2015). Leaders who execute a dominant leadership style are associated with feelings of arrogance, conceit and superiority (Cheng, Tracy, Foulsham, Kingstone, & Henrich, 2013). They use their personality skills to intimidate others.

Moreover, according to Maner (2017), they are also seen as manipulative, aggressive and also score high on dark-triad traits. Dark-triad traits include Machiavellianism and narcissism. A leader that is narcistic has a negative effect on job performance (Nevicka, Van Vianen, De Hoogh, Voorn, 2018).

When times are uncertain, a dominant leader is often more attractive than other types of leadership (Kakkar & Sivanathan, 2017). They explained that this happens because people feel that they cannot influence what happens and feel a lack of their own power. This can result in employees supporting a more dominant leader because those leaders have certain personality traits such as being agentic and authoritative. Since the corona crisis is a situation where there


is a lot of uncertainty, it would not be surprising if leaders adopt a more dominant leadership style during these times. However, it still has different effects on a follower. It is interesting to look at how it affects followers’ work stress in a stressful time, such as the pandemic.

Research has been done on leadership styles related to work stress. For example, a considerate leader contributes to lower work stress (Lopez, Green, Carmody-Bubb, Kodatt, 2010). However, there have not been studies focusing on the relationship between dominant leadership and followers’ work stress. Nonetheless, considering the personality and style of a dominant leader, a positive relationship between dominant leadership and followers’ work stress is expected. Bhatti, Shar, Shaikh, & Nazar (2010) found that leadership is also a stress factor for employees. They concluded that a relationship between a follower and a leader should be very strong. However, according to followers, their leaders do not always behave accordingly, and this results into them getting more stressed at work. It could be assumed that a dominant leader who uses fear, manipulation and aggressive tactics, can be perceived as wrong behaviour by a follower. Reviewing the literature from Bhatti, Shar, Shaikh and Nazar (2010), it can be concluded that this results in more work stress.

Considering the literature, a specific type of leadership does influence followers’ work stress. However, a direct relationship between dominant leadership and followers’ work stress has not been researched yet within the existing literature. As leadership styles show to affect work stress for followers, and personality traits of a dominant leader can be assumed to increase follower work stress, it is expected that a dominant leader increases followers’ work stress.

Additionally, this relationship has also not been investigated within the remote work environment. More specifically, it is hypothesized that dominant leadership positively relates to followers’ work stress in the remote work environment.

Hypothesis 1: Dominant leaders increase followers’ work stress in the remote work environment.

2.4 Starting a new job

Starting a new job is often exciting, but it can be scary and stressful. Now, in the middle of a pandemic, it could even be more stressful. New employers do not only have to get to know a company, its employees and their supervisor. They also have to do it all from a remote environment.

Establishing relationships between followers and leaders is more complex within the remote work environment (Avolio, 1999; Jermier & Kerr, 1997). Kelly and Kelloway (2012)


found that leadership online is not the same as face-to-face, as effective leadership is only possible in the circumstances allowing interpersonal contact with followers and leaders.

Therefore, it can be assumed that a new employee will be less likely to establish a strong relationship with their leader than employees who worked at the job before remote work was obligated. However, the results from Nelson, Basu & Purdie (1998) showed that a higher quality relationship resulted in followers being less likely to experience work stress. This would mean that employees who have a worse established relationship, in this case new employees, experience more work stress.

Another important finding of no face-to-face contact is that trust is hard to build. Mayer, Davis and Schoorman (1995) found that trust is more likely to exist in a close/face-to-face relationship than in a remote relationship between follower and leader. However, this is alarming as they also found that building trust is a crucial element within remote leadership.

Additionally, a study from Staples (2001) showed that the interpersonal trust of a follower in their supervisor resulted in higher self-perceptions of performance, higher job satisfaction and lower job stress. Lower job stress is especially interesting within this study. It could be concluded that lower interpersonal trust of a follower in their supervisor results in higher job stress. An employee who started their job before the pandemic had more time at the office to build a relationship and trust with their supervisor as they did have face-to-face contact.

Furthermore, regarding dominant leadership, it is suggested that a more dominance- based leader is more detached than a prestige-based leader (Sullivan, Landau and Rothschild, 2010). Building relationships with a more detached leader is more complex, while this is important for new employees in the remote environment to prevent more stress (Czech &

Forward, 2010).

Following the literature that a new follower cannot establish a strong interpersonal relationship with trust, especially with a supervisor who executes a dominant leadership style within the remote work environment, more work stress is expected. Therefore, I assumed that the starting point of the job moderates the positive relationship between dominant leadership and followers’ work stress. Here it is expected that the relationship is stronger when a follower started their job during the pandemic.

Hypothesis 2: The positive relationship between dominant leadership and work stress is moderated by the starting point of the job. The relationship is stronger if a follower started their job during the pandemic.


Figure 1: Conceptual model

3. Research method

Section three introduces the research method and starts with the research design. This is followed by an explanation of the sampling strategy and participants. The procedure and measurements will then be discussed, and the section finishes with the analytical plan.

3.1 Design

This research was quantitative and used a deductive approach. In this design, literature is reviewed first, and hypotheses are developed. After, data is gathered to test if the hypotheses are supported or not (Saunders, Lewis & Tomhill, 2016). Deductive research has also been chosen because it investigates a known theory and will test if it is true in situations such as the corona crisis. A survey strategy was taken regarding the research strategy, which included a self-reported online questionnaire. Last, the survey was cross-sectional. There was no consideration to change over time as the surveys were completed only one moment in time.

3.2 Sampling and participants

The sampling technique that was used was non-probability because there was a specific time frame for the survey and limited access. Within non-probability sampling, snowball sampling was used and convenience sampling. Snowball sampling was used by sending an email with the question to send it forward. Moreover, by posting the survey on LinkedIn, Facebook and WhatsApp, convenience sampling was used as the people on these sites were easy to contact and reach.

In the research, actively working adults and students were considered. This indicates that this group needed to work at least twelve hours a week, and they are required to have a direct supervisor. The supervisor is the person to whom a follower needs to report first and gets directions from. Within this study, the supervisor is regarded as the leader of a follower. People


who work at companies placed in the Netherlands were asked to fill in the survey, but they did not particularly have to be Dutch. Only participants who were working for the company for at least three months could complete the survey. This guaranteed experience with the supervisor.

The survey collected 120 filled in respondents. Twelve had to be left out because they did not finish the survey or did not consent and eleven others because they did not work remotely at all. Moreover, nine had to be removed because the participants worked less than twelve hours. There could be multiple reasons for the high number of surveys that needed to be left out. One could be that it should have been clearer that the survey should not be completed if a participant did not work remotely. Of the 88 respondents, 44 respondents were female (50%) and 44 male (50%). Their mean age was 34.08, and the standard deviation was 12.901.

The mean hours of working were 29.82 with a standard deviation of 10.432 and the mean hours that those respondents worked remotely was 26.11 with a standard deviation of 11.489 (See Appendix B1-4 for the descriptives).

3.3 Procedure

The survey consisted of multiple parts. On the first page, there was a focus on ethics within the study. It was clear what the study was for, and that participation was entirely voluntary. Everyone would be kept anonymous and was asked to give their informed consent.

During the entire survey, every participant could stop the survey at any time without a consequence. Their data would be kept private and would not be shared with others.

After giving their informed consent, participants (followers) started the survey. The survey was a self-reported online survey. During the survey multiple variables were measured.

First, respondents were asked questions about demographics to have a clear overall picture. To be able to analyse the moderator, the participants were asked if they started their job before or after the corona crisis. Afterwards, there were questions about the amount of interaction with their supervisor. Participants were then asked to fill in how many hours they worked and also how many remote. Then the independent and dependent variable was measured (See Appendix A). The survey was sent via multiple platforms as mentioned in the previous part, and data collection took place over a period of three weeks.

3.4 Measurements 3.4.1. Main variables

Dominant leadership was measured with a dominant-prestige scale by Cheng, Tracy &

Henrich (2010). The scale consisted of seventeen items measured on a 7-point Likert scale. The


dominance score is measured by averaging items 3,5,7,9,10 (reversed), 11, 12 (reversed) and 16. Originally the scale focused on the feeling of the group a follower is in. However, to get a more personal and specific idea about the direct relationship between supervisor and follower, the questions were altered from a group feeling to a personal opinion. For example, instead of using “He/she enjoys having control over other members of the group”, it was altered to “He/she enjoys having control over you”. Because of this also the Likert scale changed from 1= not at all to 7= very much to 1= strongly disagree to 7=strongly agree. Another example item is

‘He/she is willing to use aggressive tactics to get his/her way’. The scale showed sufficient reliability and was acceptable as Cronbach’s Alpha =0.929. See Appendix C1 for the reliability test of Dominant leadership.

Work stress of the follower was measured by the job-related tension index (Kahn et al., 1964). This scale was selected because it is relatively short and is grounded on a well-developed theory. The scale is a 15-item measure designed to assess job tension. By using a 5-point Likert scale ranging from 1 (strongly disagree) to 5 (strongly agree), respondents were asked how much they agree with the statements in their own situations. I also chose to use “strongly disagree” to “strongly agree” here, instead of “never bothered” to “bothered all the time”, to keep consistency and prevent participants from getting confused with the scales. An example item is: “Not knowing just what the people you work with expect of you”. The scale showed sufficient reliability and was acceptable as the Cronbach’s Alpha = .894. See Appendix C2 for the reliability test of work stress.

Last, the moderator, if a follower started their job before or during the pandemic, was measured with a multiple-choice question. The question asked when a participant started their job. The first option contained “ Before february 2020”, which indicated that they started their job before the pandemic impacted the Netherlands. The second option, “After february 2020”, indicated that the participant started its job during the pandemic.

3.4.2. Control variables

Control variables were added to rule out other possible alternative explanations. The research controlled for age, gender and hours worked per week. Age was added as a control variable as it related to work-related stress (Rauschenbach, Krumm, Thielgen & Hertel, 2013).

They found that age might affect parts of the stress process experienced at work. However, they also found that the effects were partly conflicting, which could mean that they nullify each other in the overall relation between de variables. Nevertheless, as they did find a relationship, age was added as a control variable. Age was measured with an open question, where participants were asked to fill in their age expressed in years.


Besides, gender was included as it had a relation to work stress in multiple studies.

Beena and Poduval (1992) research explained that females showed greater stress experiences.

This effect was also found by Fotinatos-Ventouratos and Cooper (2005), who found that mental and physical health was more observable amongst women than men. Trocki and Oroili (1994), who discovered the same outcome, explained this further. They found that women score higher on health issues because they are more honest and open about it on the work floor. Gender was measured with. Within the survey, participants were asked to indicate their gender in a multiple- choice question where they could choose between “Male”, “Female”, “Non-binary/third gender”, and “Prefer not to say”.

Last, work stress and hours worked per week also had a relationship in previous findings. Sparks, Cooper, Friend and Shirom (2001) researched the effects of hours of work on overall and psychological health. It suggested that work hours were positively related to ill- health. This meant that more working hours resulted in poorer health. Participants were asked about their hours worked per week in an open question expressed in hours.

3.5 Analytical plan

To analyse the data, IBM SPSS statistics version 27 was used. For the first hypothesis, the relation between dominant leadership and work stress for employees, a linear regression was used. Dominant leadership was the independent variable, and work stress the dependent variable. For the second hypothesis, the moderating effect of if the participant started their job before or after the pandemic started, on the relationship between dominant leadership and work stress, the PROCESS macro of Hays (2018) Model 1 was used. The independent variable was dominant leadership, the dependent variable work stress, and if participants started their job

before or after the pandemic began was the moderator.

4. Results

Section four discusses the results of this research. First the descriptive statistics and correlations are discussed, where the means, standard deviations and the correlation table are shown. After, the assumption checks are discussed, and the section ends with testing the hypotheses to see if they are supported or not.

4.1 Correlations

Table 1 contains all the means, standard deviations and correlations between the main and control variables. As gender is a nominal variable, it was not included in this table.

However, it is included in the regression. When p <.05, the correlation was significant. It is


shown that dominant leadership was correlated significantly with all the variables except hours worked per week. It did correlate positively and significantly with work stress. The correlation between the variables was moderate as r=0.69 with p < .01.

Furthermore, the moderator, if the participant started their job before or after corona began, correlated with all the variables. It correlated negatively and significantly with the control variables, as age (r = -.39, p < 0.01) and hours worked per week ( r = -.36, p <0.01).

However, it correlated positively and significantly with the independent and dependent variable, as dominant leadership (r=.26, p < 0.05) and work stress (r= 0.39, p < 0.01). All these correlations were very weak or weak as all the correlations had a p-value lower than .05.

Last, as mentioned, the dependent variable, work stress, correlated significantly and positively with the dependent and moderator variable. The correlation with the moderator was weak (r= 0.39, p < 0.01). However, the correlation with the independent variable was moderate (r=0.69, p < .01). It also correlated negatively and significantly with the control variables age (r=-.31, p < .01) and hours worked per week (r= -.35, p < .01). These correlations were both weak. Nonetheless, as they were significant, the control variables were included in the hypothesis testing (See Appendix D for the descriptives and correlation table).

Table 1

Descriptive statistics and correlations

Variable M SD 1 2 3 4 5

1. Ageᵃ 34.08 12.90

2. Hours worked per week 29.82 10.43 .56**

3. Start job 1.41 0.49 -.38** -.36**

4. Dominant Leadership 3.80 1.59 -.19 -.23* .26* (0.93)

5. Work stress 2.473 0.80 -.31** -.35** .39** .69** (0.89) Notes. N = 88. Cronbach's Alphas are in parentheses on the diagonal. ᵃ Age was measured in years.

* p < .05.

** p < .01.

4.2 Assumptions testing

To legitimate draw conclusions for the hypotheses, assumptions have to be tested first.

The first assumption test for the linear relationship between the independent and the dependent variable. A scatterplot was made to check this assumption. Dominant leadership was plotted


against work stress (See Appendix E1). The scatter plot showed a positive relationship between the variables. Furthermore, it showed a linear relationship which means that the first assumption was met. It was not needed to check if the residuals are independent as there is only one independent variable. We did, however, check for multicollinearity. This can lead to estimates of the coefficients that will be unstable for individual predictors

Furthermore, the third assumption considers if the residuals are normally distributed. A Normal Probability plot was made to check the third assumption. Dominant leadership was placed in the independent variable box, and work stress was placed in the dependent variable box. The plot showed that the residuals were approximately normally distributed as the blue dots are almost parallel to the black line. It was concluded that the third assumption was met (See Appendix E2).

Fourth, it was checked if the residuals were equally variable. Another scatterplot was made to check for the fourth assumption. Again, dominant leadership was put in the independent variable box and work stress in the dependent variable box (See Appendix E3).

The scatter plot showed homoscedasticity of the residuals, meaning that all the dots were distributed in the plot and not centred in the same area. Assumption four was therefore met.

Last, I checked if there were no influential data points that disproportionately affected the estimates of the regression. This assumption was tested by creating a new variable. The variable was a standardized residual which identifies outliers. With a cut-off point of (minus) two, there was one outlier. As the outlier had a value of only 2.16, it was not removed but treated with caution throughout the research.

4.3 Hypotheses testing

Firstly, the first hypothesis was tested. I tested if a dominant leader increases work stress. A multiple regression was used to test this hypothesis. The control variables, age, gender and working hours per week, were included in the first block and dominant leadership in the second block. Work stress was the dependent variable in both steps.

The R-squared of Model 2 was 51.8% (.518). This meant that 51,8% of the variance of work stress was explained by the control variables and by the independent variable dominant leadership. Furthermore, the R-squared change of Model 2 was .372 (p < .001). This meant that 37.2% of the variance of work stress was only explained by dominant leadership. Model 2 also was significant as p < .05. Moreover, the unstandardized B-value of dominant leadership was significant with a value of .319 (se = .04, t = 8.01, p < .001). This indicated that for every increase of 1-unit in dominant leadership, the work stress of an employee increased with .319.


Based on these results, hypothesis 1 was supported (See Appendix F1 for the analysis of hypothesis 1)

Subsequently, the second hypothesis was tested. Here I tested if the starting time of the job moderated the relationship between dominant leadership and work stress. I expected the relationship to be stronger when an employee started their job after the pandemic began. The PROCESS macro (model 1) of Hayes (2018) was used to test the hypothesis. The results showed that the interaction effect was not significant as the p-value was <0.05 and the confidence interval included a zero (b = 06, se = .08, t = .72, p = .47, 95% CI = -.10,.22) (See table 2). This indicates that the starting time of the job did not affect the relationship between dominant leadership and work stress. In conclusion, hypothesis 2 was not supported (See Appendix F2 for the analysis)

Table 2

Results interaction effect between dominant leadership and start job on followers’ work stress

Variable ß se t p

Constant 2.53 .40 6.40 .00

Age -.00 .01 -.64 .52

Hours worked per week -.01 .01 -1.28 .20

Gender -.05 .13 -.37 .71

Dominant leadership (x) Start job (w)

.21 .28

.14 .14

1.48 2.06

.14 .04

X * W .06 .08 .72 .47

Notes. N = 88. Dependent variable is Work stress R-squared = .54.

5. Discussion

5.1 Summary findings and meaning


This research investigated if there was a positive relationship between dominant leadership and work stress and if there was a moderating role for starting a job before or during the pandemic within the remote work environment.

The first hypothesis assumed that there was a positive relationship between dominant leadership and followers’ work stress. After analysing and evaluating the data, the hypothesis was supported. This means that a follower who experiences a dominant supervisor has more work stress than a follower who does not have a dominant leader. Moreover, the second hypothesis stated that starting a job before or during the pandemic moderates the positive relationship between dominant leadership and followers’ work stress. Assuming that starting a job during the pandemic would make the relationship stronger. After analysing the data, hypothesis two was not supported. This means that beginning a job during the pandemic did not make the relationship between a dominant leadership and followers’ work stress stronger or weaker.

5.2 Other findings and alternative explanations

The current study makes several important contributions to the existing literature. First, the research contributes to the literature as it investigates the relationship between specifically a dominant leadership style and followers’ work stress within the remote environment. Before, there have been studies on how leadership styles affect followers’ work stress in a face-to-face setting, however not explicitly on dominant leadership in a remote environment.

Results showed that a dominant leadership style causes a follower to experience more work-related stress within the remote environment. It can be assumed that these findings are in line with the assumptions made of the literature between personality traits and behaviour of a dominant leader and how it relates to followers’ work stress (De Waal-Andrews, Gregg &

Lammers, 2015; Cheng, Tracy, Foulsham, Kingstone, & Henrich, 2013; Bhatti, Shar, Shaikh

& Nazar, 2010). Additionally, it adds that there is a relationship explicitly between dominant leadership and followers’ work stress within the remote work environment. It is interesting to look at which factors of dominant leadership results in more work stress. It could be that it is because of the abuse of power or the manipulative character which stresses the followers to try their best even more which gives them stress about their tasks. However, these assumptions should be tested in future research.

Although the findings of hypothesis one were in accordance with previous findings, hypothesis two was not supported. The results showed that the effect of dominant leadership on followers’ work stress did not change when a follower started their job before or during


corona. This means that the assumed importance of building interpersonal trust and establishing a relationship with a follower face-to-face did not hold within this study. It is not necessary to question the results of previous findings because multiple researchers did find that the relationships do exist. Therefore, I looked for other explanations.

A possible reason to explain a different outcome than expected is that many sources are old. Relying on a source from twenty years ago can be tricky. The world has become more digital and electronic. This could mean that establishing a relation and building trust now is easier through the remote environment than in the past. The more significant availability of electronics and the many choices on how to communicate could make it easier to contact others.

Furthermore, another alternative relationship could be that in a pandemic, people behave differently than expected. As there were not many papers written researching these variables in times of a pandemic, no conclusions or literature could be taken from there. It could be that different leadership styles react differently than expected to a sudden crisis situation. Where, for example, the first step of building the relationship between a supervisor and a follower is that they are both in the same position where they have to adapt very quickly, which gives them a kind of empathy towards each other. From this they can later build a stronger relationship.

Now as the situation is looking better again, followers can go to the office and put effort into the face-to-face relationship. The reaction of a follower and supervisor, and the effects of this resulting from a crisis situation, such as the current pandemic, can be further explored in future research.

5.3 Limitations and future research

The current study has some limitations that need to be addressed. The most important limitation to turn to is the sampling method and size. By especially using convenience sampling, the sample is not representative for the whole population. Convenience sampling also lowers the external validity of the research and the generalizability. However, it was the most suitable method for this study as data is collected fast and cheap. Within the time frame, another sampling technique would have taken too much time or been too costly.

Furthermore, another limitation was the sample size. Before removing respondents from the data, there were 120 responses. However, many had to be deleted. One of the reasons was that some did not work remotely at all. It could have been more explicit in the message that they had to work remotely for at least some hours. A larger sample size also makes the generalizability of the results easier as more different groups are included. However, as already mentioned, the specific time frame of the study did not allow me to collect more data for a


longer period. A proposal for future research would be the same research, but with a bigger sample and probability sampling. By targeting a specific group or multiple groups, the external validity and generalizability will increase.

Another limitation is the remote work hours. Not all the participants worked fully remotely. Some also worked at the office for some hours. It was not asked if participants saw their supervisor when they were at the office, but it could influence the outcome of the variables if they did. However, since the third wave of the pandemic is ending, more people are allowed to work at the office again. Therefore, it was impossible only to target the people who work fully remotely and have a large sample size with the use of convenience sampling. Nonetheless, both means lie closely together looking at the descriptives of working hours and hours worked remotely. This indicates that the difference between working hours and hours worked remotely was not big, and participants mostly worked many hours remotely. Future researchers could focus on only remotely working employees and set them off against employees who work solely in the office. In this way, differences are most clear and give more insights on leadership and stress in different environments.

Additionally, it could be interesting to look at other leadership styles in relation with followers’ work stress within the remote work environment. Harms, Credé, Tynan, Leon and Jeung (2017) found, for example, that higher levels of transformational leadership resulted in lower stress and burnouts. Considering these findings, it would be interesting to look if this is also true in the remote work environment.

5.4 Practical implications

This study provides current and practical insights for supervisors and businesses. First, organizations should realize the consequence of a dominant leadership style for followers’ work stress within the remote work environment. It shows them that a more dominant leadership style results in more work stress for a follower. Supervisors should evaluate their leadership style and be aware that if they execute a dominant one, it causes higher work stress which can influence job performance. If supervisors become more aware that a dominant leadership style causes work stress in the remote work environment, they can alter their style to a less dominant one, resulting in less work stress for their follower that work remotely.

Furthermore, it shows companies and their supervisors that there is no difference between followers that started their job before or during the pandemic, although this was expected. This gives supervisors insights that there does not have to be a bigger focus on the employees who did start their job during the pandemic. However, as the literature showed that


highly established relationships and interpersonal trust lead to lower work stress and a higher job performance, supervisors should still focus on establishing these relationships and trust.

However, this can be done with all the employees without a focus on the ones who started their job during the pandemic.

5.5 Conclusion

Previous research found a relationship between specific leadership styles and how people feel at work. However less research was addressing this relationship, considering a dominant leadership style within the remote work environment and how it affects people who start their job during a pandemic. Given this research gap, a new research question was formulated. The study analysed if there was a direct, positive relationship between dominant leadership and work stress and if the starting moment of the job moderated this. The results showed that there was indeed a positive relationship between dominant leadership and work stress of a follower. However, this was not moderated by the starting point of their job. Besides adding more knowledge to the existing literature on remote work, the results also give supervisors executing a dominant leadership style new insights into how their style can influence their followers' work stress within the remote work environment. Evaluation of these results within their own business can strengthen the performance of working remotely.

To conclude, the research question: “Does dominant leadership positively relate to work stress of a follower within the remote work environment? And is this effect stronger when a follower starts its job during the corona crisis?” can be answered. Dominant leadership positively influences followers’ work stress. However, this relationship did not change when a follower started their job before or during the pandemic.



Avolio, B. J. (1999). Full leadership development: Building the vital forces in organizations.


Beena, C., & Poduval, P. R. (1992). Gender differences in work stress of executives.

Psychological Studies, 37(2-3), 109–113.

Bhatti, N., Shar, A. H., Shaikh, F. M., & Nazar, M. S. (2010). Causes of stress in organization: A case study of Sukkur. International Journal of Business and Management, 5, 3–14. doi:10.5539/ijbm.v5n11p3.

Cheng, J. T., Tracy, J. L., Foulsham, T., Kingstone, A., & Henrich, J. (2013). Two ways to the top: Evidence that dominance and prestige are distinct yet viable avenues to social rank and influence. Journal of Personality and Social Psychology, 104(1), 103 125. https://doi.org/10.1037/a0030398

Czech, K., & Forward, G. (2010). Leader communication: Faculty perceptions of the department chair. Communication Quarterly, 58(4), 431–457.


Dekker, D. M. (2008). Global Virtual Teams: Enhancing Effectiveness. Technische Universiteit Eindhoven, Department of Technology Management.

De Waal‐Andrews, W., Gregg, A. P., & Lammers, J. (2015). When status is grabbed and when status is granted: Getting ahead in dominance and prestige hierarchies. British Journal of Social Psychology, 54(3), 445-464.

Flores, M. F. (2019). Understanding The Challenges Of Remote Working And It ’ s Impact To Workers, 4(11), 40–44.

Fotinatos-Ventouratos, R., & Cooper, C. (2005). The role of gender and social class in work stress. Journal of Managerial Psychology.


Greco, V., & Roger, D. (2003). Uncertainty, stress, and health. Personality and Individual Differences, 34, 1057-1068. https://doi.org/10.1016/S0191-8869(02)00091-0

Harms, P. D., Credé, M., Tynan, M., Leon, M., & Jeung, W. (2017). Leadership and stress: A meta-analytic review The Leadership Quarterly, 28, 178-194.



Hon, A. H. Y., & Kim, T. Y. (2007). Work overload and employee creativity: The roles of goal commitment, task feedback from supervisor, and reward for competence. In Current topics in management, vol. 12, ed. Rahim, M. A. , 193-211. New Brunswick:

Transaction Publishers.

Howell, J. M., & Hall-Merenda, K. E. (1999). The ties that bind: The impact of leader member exchange, transformational and transactional leadership, and distance on predicting follower performance. Journal of applied psychology, 84(5), 680.

Ivancevich, J.M., Matteson, M.T. & Preston, C (1982). Occupational Stress, Type A behavior, and physical well Being. Academy of Management Journal, 25, 273-391.


Jermier, J. M., & Kerr, S. (1997). “Substitutes for leadership: Their meaning and

measurement” — Contextual recollections and current observations. The Leadership Quarterly, 8(2), 95–101. https://doi.org/10.1016/s1048-9843(97)90008-4

Kahn, R. L., & Byosiere, P. (1992). Stress in organizations. In Handbook of industrial and organizational psychology, 2nd ed., vol. 3, ed. Dunnette, M. D., Hough, L. M., 571 650. Palo Alto: Consulting Psychologists Press.

Kahn, R. L., Wolfe, D. M., Quinn, R. P., Snoek, J. D., & Rosenthal, R. A. (1964).

Organizational stress: Studies in role conflict and ambiguity. New York: Wiley &


Kakkar, H., & Sivanathan, N. (2017). When the appeal of a dominant leader is greater than a prestige leader. Proceedings of the National Academy of Sciences, 201617711.


Kelloway, E.K., Barling, J., Kelley, E., Comtois, J., & Gatien, B. (2003). Remote transformational leadership. Leadership & Organization Development Journal.


Kerr, S., & Jermier, J. M. (1978). Substitutes for leadership: Their meaning and measurement.

Organizational behavior and human performance, 22(3), 375-403.


Lopez, D., Green, M., Carmody-Bubb, M., Kodatt, S. (2011). The Relationship between Leadership Style and Employee Stress: An Empirical Study. International Journal of Interdisciplinary Social Sciences , 6(3), 170-181.


Maner, J. K. (2017). Dominance and prestige: A tale of two hierarchies. Current Directions in Psychological Science, 26(6), 526-531. https://doi.org/10.1177/0963721417714323 Mann, S., & Holdsworth, L. (2003). The Psychological Impact of Teleworking: Stress,

Emotions and Health. New Technology Work and Employment, 18, 196-211.


Mayer, R. C., Davis, J. H., & Schoorman, F. D. (1995). An Integrative Model of Organizational Trust. The Academy of Management Review, 20(3), 709. https://doi.org/10.2307/258792 Monat, A., Averill, J., & Lazerus, R. (1972). Anticipatory stress and coping reactions under

various conditions of uncertainty. Journal of Personality and Social Psychology, 24, 237-253. https://doi.org/10.1037/h0033297

Nelson, D., Basu, R., Purdie, R. 1998. An examination of exchange quality and work stressors in leader-follower dyads. International Journal of Stress Management, 5: 103 112.


Neufeld, D. J., Wan, Z., & Fang, Y. (2010). Remote leadership, communication effectiveness and leader performance. Group Decision and Negotiation, 19(3), 227-246.


Nevicka, B., Van Vianen, A. E. M., De Hoogh, A. H. B., & Voorn, B. C. M. (2018).

Narcissistic leaders: An asset or a liability? Leader visibility, follower responses, and group-level absenteeism. Journal of Applied Psychology, 103(7), 703

723. https://doi.org/10.1037/apl0000298

O’ Leary, M. B., & Cummings, J. N. (2007). The Spatial, Temporal, and Configurational Characteristics of Geographic Dispersion in Teams. MS Quarterly, 31, 433-452.

Palmer, S., Cooper, S., & Thomas, K. (2004). A model of work stress. Counseling at Work, Winter: 25.

Podsakoff, N. P., LePine, J. A., & LePine, M. A. (2007). Differential challenge stressor hindrance stressor relationships with job attitudes, turnover intentions, turnover, and withdrawal behavior: A meta-analysis. Journal of Applied Psychology 92:438-54.


Rauschenbach, C., Krumm, S., Thielgen, M., & Hertel, G. (2013). Age and work-related stress: a review and meta-analysis. Journal of Managerial; Psychology 28(7/8).


Saunders, M., Lewis, P., & Thornhill, A. (2016). Research methods for business students, Edinburgh: Pearson Education Limited.


Sparks, K., Cooper, C., Fried, Y., & Shirom, A. (2001). The effect of hours of work on health:

A meta-analytic review. Managerial, Occupational and Organizational Stress Research.

Staples, D.S. (2001), A study of remote workers and their differences from non‐remote workers. Journal of End User Computing, Vol. 13 No. 2, pp. 3‐14.


Stranks, J. (2015). Stress at work: Management and Prevention. Elsevier Butterworth Heinemann.

Sullivan, D., Landau, M. J., Rothschild, Z. K. (2010). An existential function of enemyship:

Evidence that people attribute influence to personal and political enemies to compensate for threats to control. Journal of Personality and Social Psychology, 98, 434–449. https://doi.org/10.1037/a0017457

Terry, D.J. , Nielsen, M. , & Perchard, L. (1993). Effects of work stress on psychological well-being and job satisfaction: The stress-buffering role of coworker support.

Australian Journal of Psychology, 45, 168-175.


Trocki, K.F. & Orioli, E.M. (1994), “Gender differences in stress symptoms, stress‐producing contexts, and coping strategies”, in Keita, G.P. and Jurrell, J.J. Jr (Eds), Job Stress in a Changing Workforce, American Psychological Association, Washington, DC.


Van Prooijen, J. W., & Douglas, K. M. (2017). Conspiracy theories as part of history: The role of societal crisis situations. Memory Studies, 10(3), 323–333.


Vinkers, C.H., van Amelsvoort, T., Bisson, J.I., Branchi, I., Cryan, J.F., Domschke, K.,

… van der Wee, N.J.A. (2020). Stress resilience during the coronavirus pandemic. European Neuropsychopharmacology: the Journal of the European College of Neuropsychopharmacology, 35, 12– 16.


Vrijkotte, T.G., van Doornen, L.J., de Geus, E.J. (2000). Effects of work stress on ambulatory blood pressure, heart rate and heart rate variability. Hypertension, 35.


Wang, Y., Ang, C., Jiang, Z., & Wu, C. (2019). The role of trait extraversion in shaping proactive behavior: a multilevel examination of the impact of high-activated positive


affect. Personality and Individual Differences, 136, 107-112.


Wang, B., Liu, Y., Qian, J., & Parker, S. K. (2021). Achieving effective remote working during the COVID‐19 pandemic: A work design perspective. Applied

psychology, 70(1), 16-59.


Zimmerman, B. J. (2008). Investigating Self-Regulation and Motivation: Historical Background, Methodological Developments, and Future Prospects. American Educational Research Journal, 45(1), 166–183.



Appendices Appendix A- Survey Text/Questions


In this survey I am interested in your experiences with your supervisor, his/her leadership style and what the effects of this are on work stress within the remote working environment.

Important: you need to work at least 16 hours a week and have a supervisor in order to take part in this study.

What is expected of you?

It is important that you answer honestly. There are no right or wrong answers. You can choose the answer which most closely describes you or your feelings/experiences.

Anonymity and voluntary participation

Your participation is voluntary and anonymous. Without any consequences, you can stop the survey at any time. Your responses will only be used for the purpose of my bachelor thesis in Business Administration at the University of Amsterdam and will only be seen by the researchers. If you have any questions concerning the research, please contact me: Fleur Klijn at fleurklijn@live.nl. On the consent form you only indicate whether you agree with the conditions of the study. If you have no further questions, you can now digitally sign the consent form.

Do you consent to these terms?



How old are you?


What is your gender?



Non-binary / third gender

Prefer not to say

For how long have you been working together with your supervisor?

Less than six months

Between 6 months and 1 year

Between 1 year and 1,5 year

Between 1,5 year and 2 years


Between 2 years and 2,5 years

Between 2,5 and 3 years

More than 3 years

When did you start your job?

Before february 2020

After february 2020

How often do you interact with your direct supervisor through remote channels (e.g., email, text message, phone/video call) over the past three months?

Once or twice every 1–3 months

Once or twice every month

Once or twice every week

3–5 times every week

Once or twice every day

Many times daily

How many hours do you work per week?


How many hours do you work remotely?


The following statements regard your supervisor. Please indicate the extent to which each statement accurately describes your supervisor

1. You respect and admire him/her 2. You do NOT want to be like him/her 3. He/she enjoys having control over others 4. You always expect him/her to be successful

5. He/she often tries to get his/her own way regardless of what others may want 6. You do NOT value his/her opinion

7. He/she is willing to use aggressive tactics to get his/her way 8. He/she is held in high esteem by others

9. He/she tries to control others rather than permit them to control him/her 10. He/she does not have a forcefull or dominant personality

11. You know it is better to let him/her have his/her own way


12. He/she does NOT enjoy having authority over you 13. His/her unique talents and abilities are recognized by you 14. He/she is considered an expert on some matters by you 15. You seek his/her advice on a variety of matters

16. You are afraid of him/her

17. Other do not enjoy hanging out with him/her

The following statements regard you. Please indicate the extent to which each statement accurately describes your feelings/experiences

1. Feeling that you have too little authority to carry out the responsibilities assigned to you 2. Being unclear on just what the scope and responsibilities of your job are

3. Not knowing what opportunities for advancement or promotion exist for you

4. Feeling that you have too heavy a workload‚ one that you can’t possibly finish during an ordinary work day

5. Thinking that you will not be able to satisfy the conflicting demands of various people over you

6. Feeling that you are not fully qualified to handle your job

7. Not knowing what your supervisor [senior coach] thinks of you‚ how he evaluates your performance

8. The fact that you can’t get information needed to carry out your job

9. Having to decide things that affect the lives of individuals‚ people that you know 10. Feeling that you may not be liked and accepted by the people you work with

11. Feeling you are unable to influence your immediate supervisor decisions and actions that affect you

12. Not knowing just what the people you work with expect of you

13. Thinking that the amount of work you have to do may interfere with how well if gets done

14. Feeling that you have to do things on the job that are against your better judgment 15. Feeling that your job tends to interfere with your family life

We thank you for your time spent taking this survey. Your response has been recorded.

Appendix B- Descriptive statistics B1. Descriptives gender


B2. Descriptives age

B3. Descriptives work hours

B4. Descriptives work hours remotely

Appendix C- Reliability testing

C1. Reliability testing dominant leadership


C2. Reliability testing work stress


Appendix D- Correlations


Appendix E- Assumptions testing E1. Linearity


E2. Normal Probability plot

E3. Homoscedasticity


Appendix F- Hypotheses testing F1. Regression


F2. PROCESS model 1

Run MATRIX procedure:

***************** PROCESS Procedure for SPSS Version 3.4 *****************

Written by Andrew F. Hayes, Ph.D. www.afhayes.com Documentation available in Hayes (2018). www.guilford.com/p/hayes3


Model : 1 Y : JRTI X : DL W : StartJob Covariates:

Age Gender HourWork Sample

Size: 88




Model Summary

R R-sq MSE F df1 df2 p ,7371 ,5433 ,3131 16,0621 6,0000 81,0000 ,0000



coeff se t p LLCI ULCI constant 2,5348 ,3961 6,3994 ,0000 1,7467 3,3229 DL ,2059 ,1394 1,4770 ,1436 -,0715 ,4832 StartJob ,2829 ,1375 2,0572 ,0429 ,0093 ,5565 Int_1 ,0591 ,0819 ,7216 ,4726 -,1038 ,2220 Age -,0037 ,0059 -,6397 ,5242 -,0154 ,0079 Gender -,0473 ,1286 -,3680 ,7138 -,3032 ,2086 HourWork -,0092 ,0071 -1,2825 ,2033 -,0234 ,0051 Product terms key:

Int_1 : DL x StartJob

Test(s) of highest order unconditional interaction(s):

R2-chng F df1 df2 p X*W ,0029 ,5208 1,0000 81,0000 ,4726 ---

Focal predict: DL (X) Mod var: StartJob (W)

Conditional effects of the focal predictor at values of the moderator(s):

StartJob Effect se t p LLCI ULCI 1,0000 ,2649 ,0654 4,0523 ,0001 ,1349 ,3950 2,0000 ,3240 ,0503 6,4396 ,0000 ,2239 ,4241 Data for visualizing the conditional effect of the focal predictor:

Paste text below into a SPSS syntax window and execute to produce plot.



-1,5904 1,0000 1,9243 ,0000 1,0000 2,3457 1,5904 1,0000 2,7670 -1,5904 2,0000 2,1132 ,0000 2,0000 2,6285 1,5904 2,0000 3,1439 END DATA.



*********************** ANALYSIS NOTES AND ERRORS ************************

Level of confidence for all confidence intervals in output:


NOTE: The following variables were mean centered prior to analysis:


--- END MATRIX ---




Related subjects :