1
The effects of the Interdependent Self-Construal in the
relationship between Ethical Leadership and employees’
Emotional Exhaustion and Voice
Carla Masriera Gargallo
(
11087137)Master Thesis MSc in Business Administration – Leadership and Management University of Amsterdam | Faculty of Economics & Business
Supervisor: Dr. Annebel de Hoogh
2
Statement of originality
This document is written by Carla Masriera Gargallo, who declares to take full responsibility
for the contents of this document.
I declare that the text and the work presented in this document is 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.
Table of Contents
1. Introduction ... 5
2. Theoretical Framework ... 8
Emotional Exhaustion ... 8
Ethical Leadership ... 9
Ethical leadership reducing subordinates’ exhaustion ... 11
Followers’ Voice ... 13
Self-Construal... 18
Interdependent Self-Construal ... 20
Interdependent Self-Construal as a moderator ... 21
Mediating effect of followers’ voice ... 23
3. Method ... 25
Sample and Procedure ... 25
4. Measures ... 26 Subordinates ... 26 Ethical Leadership ... 26 Emotional Exhaustion... 27 Interdependent Self-Construal ... 27 Leaders ... 27 Followers’ Voice... 27 Control Variables ... 28
3
Tenure with leader ... 28
5. Analytical Strategy... 28 6. Results ... 29 Direct Effects ... 30 Moderation Effects ... 31 Indirect Effects ... 36 7. Discussion ... 38
8. Limitations and Future Research Directions ... 42
9. Managerial Implications ... 44
10. Conclusion ... 45
11. References ... 45
12. Appendix A – Cover Letter... 53
13. Appendix B – Measures ... 55
List of figures and tables Figures Figure 1 Theoretical model 7
Figure 2 Effects of ethical leadership and interdependent self-construal on 34
emotional exhaustion. Figure 3 Effects of ethical leadership and interdependent self-construal on 35
followers’ voice Tables Table 1 Means, standard deviations, correlation and reliabilities 29
Table 2 Hierarchical regression analyses for the relationship between Ethical 31
leadership and Emotional Exhaustion. Table 3 Hierarchical regression analyses for the relationship between Ethical 31
leadership and Followers’ Voice. Table 4 Moderation analysis of Interdependent self-construal on the relationship 34
between Ethical leadership and Emotional exhaustion. Table 5 Moderation analysis of Interdependent self-construal on the relationship 35
4
Table 6 Mediation analysis of Followers’ voice on the relationship between 36
Ethical leadership and Emotional exhaustion.
Table 7 Moderation mediation analysis of Interdependent self-construal and 37
Followers’ voice on the relationship between Ethical leadership and Emotional exhaustion.
Table 8 Bootstrapping results for test of conditional indirect effects at specific 38 values of the moderator (interdependent self-construal):
Abstract
Lately, the number of studies focused on ethical leadership has increased due to the
emergence of organizational scandals and leaders influence on employees’ well-being. However, little research has been done regarding leaders’ behaviors and their impact on followers’ emotional exhaustion and their voice. I propose that ethical leadership can positively impact both of them, reducing exhaustion and increasing voice. Furthermore, I
suggest that the interdependent self-construal moderates those relationships. I found partial
support for these ideas in an analysis based on a 92 dyad survey (leaders and their
subordinates), within 5 different countries. As expected, ethical leadership was negatively related to followers’ emotional exhaustion and positively related to subordinates’ voice. The interdependent self-construal moderated the relationship between ethical leadership and employees’ emotional exhaustion. The rest of the hypotheses were not supported. Further research should focus on the independent self-construal and other followers’ outcomes such
5 Introduction
It is not a secret the worrying phenomenon which is currently taking place on the
labor market. One of the concerns of most of the companies, especially from an HR
perspective, is the so called absenteeism. This issue, a part from being related to employees
well-being (Warr, 1999), can also have a dramatic impact at different levels of the
organization, such as loss of productivity (Johns, 2011). In addition, the main focus of recent
studies in the causes (Blau, & Boal, 1987; Mowday, et al., 2013; Schaufeli et al., 2009) and
consequences (Johns, 2011; Ybema et al., 2010) of this concept, suggests that it may actually
be relevant xxxx and deserves more attention. For instance, Schaufeli et al. (2009) have
found that the antecedent of burnout predicts future absence duration. Other authors have
demonstrated that emotional exhaustion, one of the three dimensions of burnout (Maslach,
Schaufeli, & Leiter, 2001), is one of the principal causes of medical absences (Michie &
Williams, 2003; Toppinen-Tanner, Ojajärvi, Väänaänen, Kalimo, & Jäppinen, 2005).
Whereas some literature suggests that leaders may influence employees’ well-being (Donaldson‐Feilder, Munir, & Lewis, 2013; Liu, Siu, & Shi, 2010; Arnold et al., 2007), little
research has been done regarding leaders’ behaviors and their impact on followers’ exhaustion, particularly in relation to the role of ethical leadership. This kind of leaders,
perceived as integral, honest and trustworthy figures, are characterized for acting according
moral and proper values (Brown, Treviño, & Harrison, 2005). They are also known for
considering employees’ interests and for supporting them, thus building a relationship based
on trust and reciprocity. This concern for subordinates and commitment to them is a powerful
leadership tool, which has the ability to comfort employees and it may reduce their sources of
anxiety and chances to feel exhausted.
Furthermore, Brown and Treviño (2006) have learned that ethical leaders encourage
6
as satisfaction, motivation or organizational commitment. In addition, Brown, Treviño and
Harrison (2005) have found that ethical leaders stimulate their followers to work harder and
to report any problems they perceive within the organization. However, this characteristic of
ethical leadership and its effects, which involves a particular employee behavior designated
as follower’s voice, has not been deeply analyzed. The concept of voice refers to workers’
willingness to speak up and to propose alternatives to improve a problematic situation (Van
Dyne & LePine, 1998). This employees’ proactivity requires our attention because this
initiative could not only benefit the individuals, but it could also make a difference within
co-workers relationships and in the performance of the entire organization (Morrison, 2011).
The practice of an ethical leadership style is probably not enough to avoid the
emergence of this psychological strain or to stimulate employees’ voice. Furthermore,
follower traits surely play an important role in the interplay. One potential, important
subordinate characteristic that may impact the effectiveness of ethical leadership is the degree
of followers’ interdependent construal. This more collectivistic dimension of the
self-construal has its origins in non-Western societies, and emphasizes the necessity of individuals
to socialize, share, belong and connect to their environment (Markus & Kitayama, 1991). As
a result of this, and according to social identity theory (SIT) (Tajfel & Turner, 1979), employees’ needs and their perceptions of similarity with their supervisors will lead them to connect and be more open to their supervisor. Therefore, the higher social identification
employees demonstrate in relation to their ethical leaders, the stronger the impact may be of
their management on exhaustion and voice.
Finally, the two aforesaid direct effects of ethical leadership on followers’ voice and
emotional exhaustion are not the only relationships that can be found between these variables. Followers’ voice may also be mediating the relationship between the other two. The positive outcomes employees experience when they exteriorize their feelings and thoughts, such as an
7
increase of their sense of control (Morrison, 2011) and a higher perception of fairness and
equity (Folger, 1977; Kanfer et al., 1987), suggest that followers’ voice will have an impact on employees’ emotional exhaustion.
Using a sample of 92 leader-employee dyads, I will focus my research on the effect of
ethical leadership on a subordinate’s emotional exhaustion and voice, and the role that the
interdependent self-construal has on both relationships. I will also analyze the mediation
effect that followers’ voice has on the relationship between ethical leadership and
subordinates’ exhaustion (see Figure 1 for the theoretical model). The findings of this study
will extend the existent knowledge on ethical leadership effects and, hopefully, with such
awareness we then will be able to reduce some work stressors, which are the reason for a
wide range of negative effects such as depression or anxiety.
8 Theoretical Framework
Emotional Exhaustion
Emotional exhaustion occurs when individuals experience feelings of frustration and
saturation because they do not have enough resources available, such as time, tools or
external support, to cope with some demands (Maslach, Schaufeli, & Leiter, 2001). Being
one of the three dimensions of job burnout, preceding cynicism and inefficacy, exhaustion is
considered the most important individual stressor and it has been the most analyzed, probably
because it can be scrutinized without considering the other two.
The three dimensions of burnout are measured together through the MBI-General
Survey. MBI-GS was elaborated to reach a wider amount of population after the development
of two previous assessments exclusively focused on health care and education, MBI-Human
Services Survey and MBI-Educators Survey, respectively. Usually, this appraisal has been
conducted not only to study the findings related to work outcomes, but also with employees’
well-being (Maslach, Schaufeli, & Leiter, 2001). Furthermore, research on the variables
which impact each particular dimension has received especial attention and it has been
divided into two broad categories: demands and resources. Previous studies have postulated
demands as the main predictors of emotional exhaustion, while resources have been
considered to be more associated with depersonalization or personal accomplishment (Lee &
Ashforth, 1996).
In theirmeta-analytic examination, Lee and Ashforth (1996) found that five of the eight
demands analyzed (role ambiguity, role clarity, role conflict, stressful events and workload),
and considered job stressors, were strongly associated with the emergence of exhaustion.
From the resources approach, support provided from the supervisor or a coworker was also
9
cohesion, autonomy (as a form of job enhancement opportunity) and unmet expectations (as a
contingent reinforcement) were also contemplated as influencers of exhaustion.
For instance, social support results to reinforce the proposition that a weak relationship
between leaders and their subordinates increases the probability of experience psychological
distress. Another type of characteristic which seems to stimulate burnout are those related to
professions which involve a component of sensibility and consideration, either for some tasks
or for other individuals, and that require emotional demands. Finally, the organizational
framework can also be a potential influence, especially when some implicit agreements, such
as fairness and justice, are disrupted (Maslach, Schaufeli, & Leiter, 2001).
The existent difference between countries is a further contextual factor that has been
analyzed. Using the MBI as a basis, studies have shown that levels of emotional exhaustion
are higher in North-America than in Europe. Researchers suggest that cultural mind-sets
could be one of the reasons for this variance, although it has not been proved. For example,
an individualistic posture could promote the appearance of the dimension of cynicism, while
a more collectivistic mentality could develop higher levels of emotional exhaustion
(Golembiewski et al., 1996; Maslach, Schaufeli, & Leiter, 2001).
Ethical Leadership
Brown et al. (2005) described ethical leadership as a behavior through which leaders
should act and interact with others, especially their subordinates, in a proper and upright way.
Since then, a lot of studies have been conducted arousing due to a recent sequence of
scandals in organizations, where the blame is usually on top managers and which are the
principal focus of these researches. Empirical evidence suggests a direct relation between these leaders’ actions and negative consequences for the company, such as frauds. However, other outcomes including role modelling, subordinate perceptions and interpersonal
10
relationships, which can be precedents of this unethical behavior, could be further
investigated.
Fundamentally, a leader has to present two properties in order to be considered
ethical. The first one is incorporated in the definition of moral person, and it is related to the
internal dimension of the leader. It refers to the conduct that individuals have to develop
according certain values, but also in how their actions are perceived by others. The main
qualities of a moral person can be classified in three different categories: traits, behaviors and
decision-making. Integrity, honesty and trustworthiness are contemplated as expected traits of
ethical leadership, whereas correct behavior, openness, personal morality and concern for
people can be understood as behaviors. Additionally, ethical leaders should base their
decisions according to their principles, and being impartial, fair and just (Treviño, Hartman,
& Brown, 2000).
The second attribute, implied in the definition of moral manager, involves external
parties, mainly the subordinates, and aims to the effects that the leader has on them through
being a role model and depending on the quality of their relationship. In order to be influent
and to achieve the purpose sought with their modelling, ethical leaders should act smartly,
maximizing the signaling effect of positive actions and reducing those that may be harmful.
They also can guide employees, rewarding and punishing the behaviors they like or dislike,
respectively. Last, they should clearly convey ethical and moral messages, to raise awareness
of morality across the organization (Treviño, Hartman, & Brown, 2000).
Probably, due to all of these distinct characteristics and the broad scope covered by
ethical leaders, this leadership style has largely been compared with other kinds of leadership
and constructs that can overlap. For example, Bass and Avolio (1993) have identified an
ethical element in idealized influence, one dimension of transformational leadership. In both
11
for their employees. Other studies have shown the impact of leader’s honesty, trustworthy
and integrity on leadership effectiveness (Den Hartog, House, Hanges, Ruiz-Quintanilla, &
Dorfman, 1999; Kouzes & Posner, 1993; Posner & Schmidt, 1992), also key components of
idealized influence. Finally, ethical leadership has been associated to employees’
consideration and fair treatment. Attributable to their authority and impact on substantial
decisions, ethical managers need to be aware of employees’ perceptions and judgements of
their acts, showing them a high degree of decency and respect (Brown et al., 2005).
All the aspects mentioned above are the basis for the Ethical Leadership Scale (ELS)
build by Brown et al. (2005). The actual criterion consists of ten items that incorporate
several features in its definition, which include a wide range of dimensions and reflect the
connection between ethical leadership and other constructs. For instance, the instrument
analyses both managers’ individual behavior but also their consideration and preoccupation
for their employees. This measure has been proved as a valid predictor for significant
followers outcomes such as organizational commitment, voice behavior, job satisfaction,
satisfaction with the leader, perceived leader effectiveness, job dedication, follower
willingness to report problems and psychological safety (Brown et al., 2005; Brown &
Mitchell, 2010; Walumbwa & Schaubroeck, 2009), among others. However, the effect that
ethical leaders have on employees’ emotional exhaustion has not received much attention yet.
Ethical leadership reducing subordinates’ exhaustion
In line with the aforementioned positive outcomes of ethical leadership, and
especially regarding leaders’ impact on employees’ behaviors, I hypothesize that this kind of
leadership may be relevant in reducing subordinates’ emotional exhaustion. Through different mechanisms, ethical leaders should be able to manage employees’ sources of stress, avoiding consequent situational strains.
12
To begin with, the medical field has probably been one of the most used to study
different antecedents and mediators of emotional exhaustion. For instance, Stordeur et al.
(2001) tested how transformational and transactional leadership influence emotional
exhaustion levels of nursing staff. They found that work stressors had an important role in
this relationship. Nevertheless, it seems that few empirical studies refer to ethical leadership
and its impact on employees’ exhaustion, at least to my knowledge, and that there are other
variables that may drive this relationship.
De Hoogh and Den Hartog (2008), in their multimethod study of ethical and despotic
leadership, acknowledge the concept of role clarification. They define it as “leader's
transparency, engagement in open communication with followers and clarification of expectations and responsibilities so that employees are clear on what is expected from them” (p. 298). This clarity should reduce confusing situations and some job stressors such as role
conflict and role ambiguity, which can foster the appearance of emotional exhaustion (Lee &
Ashforth, 1996).
Moreover, Kalshoven, Den Hartog and De Hoogh (2011) use this dimension of ethical
behavior, among others, to build the Ethical Leadership at Work Questionnaire (ELW). The
ELW permits to distinguish the different factors by which employees are influenced by
ethical leaders. In particular, this construct includes a variety of items that refer to
performance expectations, and clarification of priorities and responsibilities. In their study,
Kalshoven, Den Hartog and De Hoogh (2011) found that role clarification was related to trust
in leaders and their effectiveness. Thereby, these results suggest that ethical supervisors will
try to communicate their messages in a clear and objective way, in order to reduce any chance of employees’ mistrust and/or discomfort.
Another workforce demand that can be decreased through ethical supervisors is role
13
qualitative saturation, provoked by employees’ lack of skills, can explain their difficulties in achieving targets. On the other hand, a quantitative overload can also lead to employees’
frustration due to a shortage of resources (Cordes & Dougherty, 1993). Ethical managers
should rely on their tacit nature and sincere concern for their subordinates, and provide them
with enough resources, both psychological and material, in order to avoid these situations.
Socially supporting them, ethical leaders may optimize employees’ well-being; providing
them with sufficient resources and reducing the pressure they may be exposed to (Kalshoven
& Boon, 2012).
Finally, two last considerations should be taken into account. On one hand, ethical
leaders are considered honest, fair and sponsors of equity (Brown et al., 2005; Treviño,
Hartman, & Brown, 2000). On the other hand, the properties of perceived fairness and
procedural justice, both of which are related to organizational framework and leadership
behaviors, have the ability to enhance psychological safety of employees and, therefore, to
reduce their exhaustion. For instance, the other possible cause which could lead to emotional
exhaustion is also controlled by ethical leaders, and the first supposition is as it follows:
Hypothesis 1. Ethical leadership is negatively related to followers’ emotional exhaustion.
Followers’ Voice
One of the responsibilities of a leadership position is to have repercussions on the
organization they work for through employees’ behaviors. Besides the already explained and
detailed effect on employees’ emotional exhaustion, ethical leadership can also have further
important impact on followers’ conducts. Some outcomes previous researched are those related to subordinate’s ethical decision-making, prosocial and counterproductive behavior, work attitudes such as satisfaction, motivation or organizational commitment (Brown &
14
This last variable, willingness to report problems, is related to followers’ voice extra-role behavior. Employees’ voice, considered part of the challenging dimension of the
organizational citizenship behavior (OCB) (Van Dyne & LePine, 1998), has often been
analyzed from a contextual performance approach and it can be categorized into two broad
streams. The first one is related to the influence that an individual has on a decision-making
procedure. The second one emphasizes workers predisposition to suggest improvements
when they identify obstacles. In their study about the multidimensional constructs of silence
and voice, Dyne, Ang and Botero (2003) propose three different types of voice determined by employees’ incentives. Acquiescent voice is based on resignation and has an important component of disengagement with the organization and/or other individuals. Defensive voice
builds on fear and acts as a shield, being a tool of self-protection. Last, prosocial voice arises
from cooperatives behaviors and has a proactive factor which highly correlates with the OCB
measure. Because the scope of this study is on positive outcomes of ethical leadership, I will
focus on the measurement of prosocial voice.
Furthermore, three different levels have been identified to explain the reasons why
some people tend to engage in this behavior more than others. The first one emerges from one’s own nature and signal to individual differences. LePine and Van Dyne (2001) proposed and tested a model which supported their propositions on the affects that the big five
characteristics have on voice behavior. In summary, they found out that conscientiousness
and extraversion promote this conduct, whereas neuroticism and agreeableness inhibit it. The second one is related to employees’ perceptions of the organization’s management, and it points out the necessity of workers to speak up when they are discontented and search for
some changes (Rusbult, Farrell, Rogers, & Mainous, 1988). Finally, the third one focuses on
behaviors influenced by the environment one is exposed to. This dimension suggest that,
15
raise of employees’ voice, they will not do so if they distinguish an hostile situation or if they
encounter a lack of support (Milliken, Morrison, & Hewlin, 2003).
Generally, employees’ voice is considered to be more beneficial than silence is, and a
great variety of voice outcomes have been recognized. From an organizational perspective,
for example, its relevance on effective decision making and detection of mistakes has been
proved as significant. Moreover, the sharing of knowledge within groups and the resulting
improvement on learning and performance is considered another benefit of this behavior.
From an individual perspective, voice has the ability to increase employees’ sense of control
and autonomy, enhancing their satisfaction and motivation and reducing their levels of strain.
Furthermore, the verbalization of one’s perturbations may have a favorable impact on one’s
well-being. Unfortunately, there also exist some negative aspects that need to be taken into
account. Distorted colleagues perceptions can harm one’s public image. Co-workers may
confuse others’ recommendations with criticism, which would bring them to designate these
employees as informers. Finally, suggesting further adjustments may lead to different kinds
of punishment such as low performance appraisals and consequent lack of promotions
(Morrison, 2011). Thus, the decision to speak up will probably be dependent on the
evaluation of pros and cons, and it only will be carried out when positive consequences
exceed the negative ones.
Ethical leadership enhancing followers’ voice
The significant impact of ethical leaders on employees’ outcomes has been supported
by different studies. Ethical leadership can influence followers’ perceptions and their
consequent feelings and behaviors, such as psychological safety, job satisfaction and
willingness to report problems (Brown et al., 2005; Brown & Mitchell, 2010; Walumbwa &
16
A research conducted by Zhu et al. (2015) has found that ethical leadership is positive related to followers’ voice behavior through social identification mechanisms (relational identification and organizational identification). Other authors, for example Avey et al.
(2012), have suggested that ethical leaders encourage followers’ voice through two different
means. On the one hand, employees tend to copyethical leaders behaviors which are seen as
a role model. Therefore, and according to ethical leadership definitions (Brown et al., 2005;
Treviño, Hartman, & Brown, 2000), ethical leaders will probably speak up when they
disagree with others or observe unethical behaviors, and their subordinates will emulate them.
On the other hand, these kinds of managers provide employees with the support necessary to
voice their concerns and opinions. Moreover, the results of this study support the mediating role of employees’ voice behavior in the relationship between ethical leadership and employees’ psychological well-being.
Participation on decision making procedures, one of the voice streams distinguished
by van Dyne and LePine (1998), may also be influenced by ethical leaders behaviors (Brown
et al., 2005). As it has been exposed earlier, ethical managers are characterized for having
high standards of justice and for leaning on fair decisions (Treviño, Hartman & Brown,
2000). Due to this mentality, these leaders may give employees the opportunity to participate
in these processes and they would probably consider all of their ideas, suggestions and points
of view, before making the final decision.
The other stream of employees’ voice, people predisposition to suggest improvements (van Dyne & LePine, 1998), can also be reinforced by ethical supervisors. In order to be
considered ethical, this kind of leaders should support their subordinates, taking into account
their needs and fostering a safe environment, completely avoiding any kind of hostility. This,
in turn, should encourage employees to raise their hand and to point out worrying issues with
17
In most of the cases, employees need to feel confident before speaking up. As Milliken,
Morrison and Hewlin (2003) stated, employees will not engage in this behavior if work
atmosphere is hostile or support is insufficient. Indeed, in their empirical study about the
mediating role of ethical leadership and psychological safety, Walumbwa and Schaubroeck
(2009) found that perceptions of this security mediated the relationship between ethical leadership and followers’ voice. For instance, we can suggest that ethical leaders have the potential to generate the required fair climate to encourage employees to act, and to provide
them with sufficient resources to develop their prosocial voice.
This relationship can also be explained using the concept of reciprocity. From a social
exchange approach, Blau (1964) suggests that a high-quality relationship, based on trust and
benevolence, will lead to reciprocated exchanges between the agents involved. A particular
perspective to address this phenomenon is established between employees and their
supervisors, and the moral obligation subordinates demonstrate taking responsibilities that go
beyond their contract (Liden & Graen, 1980; Wayne & Green, 1993). For example, on an
empirical research, Settoon, Bennett and Liden (1996) confirmed that relationships based on
mutual trust stimulate favorable employees’ conducts such as extra-role and citizenship
behaviors. Therefore, the relationship developed from an ethical leader – follower dyad will
probably be built upon high standards, inciting employees to reciprocate their managers and
repaying them back for the consideration received. Subordinates may contemplate the option
to, in a courteous manner, suggest modifications and new ways of doing things and interact,
in order to improve individual, team and organization performance. And all together leads to
the following proposition:
18
Self-Construal
External influences such as ethical leadership are not the only factors that can affect employees’ behaviors. One’s own characteristics can be significant as well. The concept of self-construal, which has its origins in Markus and Kitayama (1991) article of the Self, can be
considered another possible explanation. The authors build their framework on a cultural
context and use the distinction of Western and non-Western societies to explain the main
differences on the identities of the individuals and the conceptions about the self. First
definitions clearly delimited both dimensions depending on cultural antecedents, although
further studies suggested that individuals can develop both of them. They can either occur
simultaneously, or emerge in specific situations or points in time (van Baaren et al., 2003).
On the one hand, and from a Western approach, a person who develops the
independent self-construal is defined as "an independent, self-contained autonomous entity
who (a) comprises a unique configuration of internal attributes (e.g., traits, abilities, morals,
and values), and (b) behaves primarily as a consequence of these internal attributes" (Markus
& Kitayama, 1991, p. 224). On the other hand, and from a non-Western perspective, the
interdependent self-construal is based on the idea that “experiencing interdependence entails
seeing oneself as part of an encompassing social relationship and recognizing that one's
behavior is determined, contingent on, and, to a large extent organized by what the actor
perceives to be the thoughts, feelings, and actions of others in the relationship” (Markus &
Kitayama, 1991, p. 227).
In order to understand its functioning, previous research has been focused on
explaining the causes of self-construal differentiation. Marcus and Kitayama (1991)
compared the independent and the interdependent view of the self to suggest that different
construals may induce to cultural differences. Cross and Madson (1997) reviewed the
19
approaches between men and women regarding cognition, self-related motivations, emotions,
and relationships.
Other authors have analyzed the influence of the self-construal on direct coping
strategies and levels of stress (Cross, 1995), and whether differences in the self-construal
mediate the relationship between ethnicity and coping (Lam & Zane, 2004). In a longitudinal
project based on data collected among undergraduate students, Cross (1995) found that the
level of stress among international students was reduced due to the predominance of the
independent self-construal, positively related to direct coping strategies. Contrary, the levels
of stress experimented by Asian students was caused by the interdependent self-construal.
Lam and Zane (2004) extended the research on coping strategies and the self-construal, and
found that the dominant independent self-construal in White Americans may support the
adoption of primary control technics. Opposed, the dominant interdependent self-construal in
Asian Americans partially mediated the relationship between ethnic differences and
secondary control.
Moreover, van Baaren et al. (2003) examined the moderation effect of self-construal
orientation on mimicry, and found that interdependent self-construals lead to lower levels of
nonconscious mimicry compared to independent self-construals. The experiment was
replicated using an American sample as the independent self-construal subjects, and a
Japanese sample as the interdependent self-construal subjects. The results were statistically
similar.
In addition, Cross et al. (2003) focused on the potential outcomes of different
self-construal such as social behavior and well-being, and the results showed that the relationship
between consistency and well-being was stronger for individuals with a low relational
self-construal than for those individuals with a high relational self-self-construal. Finally, Hackman et
20
factors of the independent and the interdependent self-construal, instead of assuming that
they are both part of the same construct. Second of all, the concept of the self-construal can
be a predictor of leadership communication style.
When we talk about ethical leaders, we usually refer to the consideration they have
toward others and how is the quality of their relationship (Brown et al., 2005; Treviño,
Hartman, & Brown, 2000). We also acknowledge the important role of reciprocity and social
relationships. For instance, a recent study from Zhu et al. (2015) has focused on the mediator role of followers’ relational identification (Walumba & Hartnell, 2011) on the relationship between ethical leadership and followers’ voice to determine how followers’ identification with the leader influences their outcomes. As expected, they found that ethical leadership was positively related to followers’ voice. In the same research, they also used the concept of morality beliefs to explain why followers’ behaviors vary depending on how and upon what
extend they identify with their leaders and internalize their beliefs. The authors suggest that
those individuals with high entity moral beliefs will engage in a stronger relationship with
their leader and, thereby, they will be more willing to speak up. Again, the results confirmed
the expectations. This relational identification and the concept of moral beliefs clearly have
more common aspects with the interdependent self-construal rather than with the independent
self-construal. Therefore, in this study, I will only analyze the interdependent self-construal
and its impact on the relationship between ethical leadership and subordinates’ emotional
exhaustion and voice.
Interdependent Self-Construal
The interdependent self-construal is defined as a “flexible, variable self that emphasizes
(a) external, public features such as statuses, roles, and relationships, (b) belonging and fitting in, (c) occupying one’s proper place and engaging in appropriate action, and (d) being
21
indirect in communication and reading others’ minds” (Singelis, 1994, p. 581). Japanese culture is the benchmark of the interdependent self-construal. However, Asian, African,
Latin-American and Southern European cultures are also categorized in this dimension.
Collectivism is part of their nature, as well as socialization, connectedness, belongingness
and reciprocity. Individual features of one are not significant anymore, and relationships with
others become the most valuable asset. People’s self-realization and satisfaction stem from
the recognition of their fit in a social context, where individuals identify themselves similar to
others, and where a sense of interdependence is salient.
The predominance of an interdependent self-construal does not mean that inner traits are
inexistent, but they are contingent to the situation. In other words, their value and
significance depend on each circumstance and, compared to relational features, they are
powerless. This behavior is often voluntary; individuals are mature enough to prioritize and
to decide when the best moment to display their deepest feelings is. Unlike the independent
self-construal, the interdependent cannot be defined as a static and specific construct. Instead,
its configuration is a combination of personal traits, situations and individuals’ development
through their interactions (Markus & Kitayama, 1991).
Interdependent Self-Construal as a moderator
Recently, the study of the self-construal has progressed and some researchers have
identified its impact on cognitive and emotional processes, especially in social and
organizational contexts. These new insights are often well supported by other theories, which emphasize followers’ identification. One of the most frequently used, the Social Identity Theory (SIT) (Tajfel & Turner, 1979), describes how individuals perceive themselves as
similar or different than others, and how they decide to become part of a particular collective.
22
related to one’s own characteristics, and the social identification, which stresses the need to fit in a group (Ashforth & Mael, 1989).
The social identification dimension has been associated with the collective self-construal,
due to the importance of belongingness in both of them. For example, Brewer and Gardner
(1996) use their personal situation as social psychologists to explain the core of social
identity and how they are more predisposed to ascribe themselves attributes they share with
other members in the group, and which make them different from other collectives. This
differentiation between intragroup differences, intragroup similarities and intergroup
differences is useful to explain the prevalence of the personal or collective self (Simon,
Pantaleo, & Mummendey, 1995). In other words, the social identity theory supports the
assumption that individuals, employees in this case, high on the social identity dimension
makes them more likely to identify with another person with the same values, the leader.
In the earlier description of an ethical leader, the dimension of moral person implies that
the leader should believe in certain values and the subordinate should perceive them as the
same. According to social identity theory, followers also hold particular values (personal
identity), which may be similar or different from their leaders’. The dimension of social
identification suggests that subordinates with the same values of their leaders, will be more
inclined to engage in a relationship with them. In this situation, individuals with a high
degree of interdependent self-construal will decide to boost this connection and to take
advantage of its synergies.
Moreover, in a review of the self-construal, Cross, Hardin and Gercek-Swing (2011)
describe both dimensions from three different perspectives: cognitive, affective and
motivational. The independent self-construal makes individuals unaware of the context,
23
values. Contrary, the interdependent (or collective) self-construal increases personal
consciousness of the environment, expecting engaging behaviors and allowing self-criticism.
This categorization means that, an individual with the interdependent self-construal
developed will have more chances to identify with an ethical leader rather than an individual
who is characterized for the predominance of the independent self-construal. Furthermore,
the collective self-construal will promote subordinates to engage in a closer relationship with
their leaders, and they will be able to benefit from ethical managers and their consideration
for them, for example, in terms of their well-being and psychological safety. Therefore:
Hypothesis 3. The negative relationship between ethical leadership and followers’ emotional
exhaustion is moderated by followers’ interdependent self-construal, so that this relationship
is stronger for higher levels of interdependent self-construal.
Hypothesis 4. The positive relationship between ethical leadership and followers’ voice is
moderated by followers’ interdependent self-construal, so that this relationship is stronger for higher levels of interdependent self-construal.
Mediating effect of followers’ voice
Finally, the two aforesaid direct effects of ethical leadership on followers’ voice and
emotional exhaustion are not the only relationships that can be found between these variables.
There is an indirect link existing among the three that is also worth to be considered; the mediating role of followers’ voice on the relationship between ethical leadership and subordinates’ exhaustion.
One of the individual outcomes of follower’s voice mentioned above is the increment of employees’ sense of control, which can result to less stress. Parker (1993), in her empirical study using a sample of hospital nurses, demonstrated that perceived high levels of control
24
result into high levels of personal well-being. The author based her hypotheses on the
premise stated by Bandura (1986) that discretion seemed to minimize stressful situations
because its influence on employees’ self-confidence. Ethical managers, in order to satisfy subordinates’ needs, will allow them to participate in decision-making procedures, intensifying their feeling of control (Bordia et al., 2004). This permission, in turn, will lead
employees to feel more supported by their supervisors and they will not be saturated because
a lack of resources. This perception of having control over issues which concern them, will
also reduce job stressors and individuals will feel less exhausted.
Another singular outcome that has been recognized in earlier sections for influencing
emotional exhaustion is the perception of fairness and equity. Folger (1977), in his research
focused on distributive and procedural justice, concluded that giving the subjects the
opportunity to express their preoccupations would only lead to positive reactions if managers
reciprocated them. In other words, when allowing employees to participate, supervisors
should take into account their opinions and act accordingly their specific requests. Once
considered listened and being satisfied their expectations, workers will perceive a high degree
of fairness and their levels of stress will be reduced.
In this line, Kanfer, Sawyer, Earley, and Lind (1987) also emphasize the relationship
between participation, perceptive fairness and employees’ attitudes and behaviors. In their
initial literature review, the authors mention the allowance of voice as a signal of procedural
fairness. Besides the recurrent analyzed participative decision making, they also discuss
employees’ contribution in goal-setting, enhancing their job satisfaction and motivation. In
both cases, the communication between supervisors and subordinates plays an important role. Again, the reciprocating conduct based on employees’ proposals and respective managers’ consideration, will determine followers’ judgements about the integrity of the system. The results of their investigation confirm the impact of participation voice in procedural and
25
distributive justice, and its succeeding influence in task performance. According to these
arguments, it is suggested that:
Hypothesis 5. The positive relationship between ethical leadership and followers’ exhaustion
is mediated by followers’ voice.
Hypothesis 6. Ethical leadership is related to followers’ exhaustion via conditional indirect
effects, such that the interaction between ethical leadership and followers’ interdependent self-construal is related to followers’ voice, which in turn is related to followers’ exhaustion.
Method
Sample and Procedure
The total sample of this study consists of 92 dyads of different labor markets and
nationalities such as The Netherlands, Greece, Spain, Germany and Belgium. Four sets of
questionnaires were prepared, two in Dutch and two in English, to administer to
supervisor-subordinate dyads, working more than 20 hours per week. The surveys were designed to
carry out a wider project in charismatic and ethical leadership, which involves six students of
the Leadership and Management track of the Business Administration master degree at the
University of Amsterdam. A broad range of variables, required for each individual research
model, were included in the questionnaires. The Online software, Qualtrics, was used to
design the surveys. They were later distributed through E-mail addresses (to supervisors,
subordinates, or both) with the support of a cover letter (Appendix A), emphasizing the
guarantee of anonymity and promoting honesty. The total of surveys administered was about
174 dyads; 144 subordinates and 122 managers started them, but only 116 and 108,
respectively, finished it. After removing the incomplete surveys, missing values, and
26
On one hand, among the leaders, there were 62.2% males and females 37.8%. A high
school diploma was the highest educational level obtained by 7.7% of these leaders, while
4.4% had received technical education, 54.5% were college graduates and 33.6% had a
Master degree. They were on average 44.86 years old and they had been working with the
organization around 16.50 years. The relationship with the employee matched was 3.47 years
on average, and they used to contact each other on a daily (79%) and weekly (21%) basis.
On the other hand, among the employees, 28.7% were men and 71.3% were women.
Regarding their highest educational level, 6.3% were graduated from high school, 7% had
taken technical education, 58.7% were undergraduates and 28% hold a Master degree. On
average, they were 36.21 years old and their tenure with the organization was about 1.78
years.
Finally, 1.4% of the respondents were from Belgium, 3.5% from Germany, 23.8%
from Greece, 61.5% from The Netherlands and 9.8% from Spain. Therefore, 61.5% of the
surveys were provided in Dutch, the remaining 38.5%, in English.
Measures
All of the four constructs analyzed were measured through validated 7-point Likert
scales (1 = strongly disagree – 7 = strongly agree).The participants were asked to what extent
they agree or disagree with a specific statement.
Subordinates Ethical Leadership
Ethical leadership was assessed by the “Ethical Leadership Questionnaire” developed by Brown, Treviño and Harrison (2005). The ELS, with a Cronbach’s Alpha of .92, measured
27
and some of the statements, for example, were: “Listens to what employees have to say” or “Makes fair and balanced decisions”.
Emotional Exhaustion
Emotional exhaustion was measured through 6 items based on the Maslach Burnout
Inventory (Maslach & Jackson, 1986) (α = .90) and Schaufeli and van Dierendonck (1995). “I
feel mentally exhausted by my work” and “A working day is a heavy duty for me” are two of
the statements asked. This measure was considered self-reported, and inserted on the subordinate’s questionnaire.
Interdependent Self-Construal
The interdependent self-construal was examined through 14 items based on
Gudykunst and Lee (2003) with the following alphas: United States = .80, Japan = .84, Korea
= .85, and Australia = .85 (Gudykunst et al., 1996). This construct was also introduced in the subordinate’s survey using several items such as: “I consult others before making important decisions” and “I give special consideration to other’s personal situations so I can be efficient
at work”.
Leaders
Followers’ Voice
Finally, 6 items based on Van Dyne and LePine (1998) (α=.94) were used to analyze subordinate’s voice. “This particular co-worker develops and makes recommendations concerning issues that affect this work group” and “This particular co-worker speaks up in
this group with ideas for new projects or changes in procedures” were two of the statements used in the supervisor’s questionnaire.
28
Control Variables Tenure with leader
According to the whole model of this study, and due to the potential effect that
familiarity and closeness may have in the relationship between leaders and their subordinates,
employees´ tenure with leader was included as a control variable.
Analytical Strategy
The results of the surveys conducted through Qualtrics were cleaned, matched and
processed on the software package SPSS Statistics. After removing the incomplete surveys
with missing values, the hypotheses were tested on a data base of 92 leader-employee dyads.
Demographic variables were used to calculate descriptive statistics of respondents and it was
assured the nonexistence of counter indicative items, therefore there was no need to recode. A
scale reliability test was also computed to guarantee reliability of the constructs. All the
scales used appeared to be good on reliability, all of them with Cronbach´s Alpha above .70.
The corrected item-total correlations reflected the good correlation between all the
items and the total score of the scale (>.30), except for two particular items on the
interdependent self-construal construct such as “I consult with co-workers on work-related matters” (.29) and “My relationships with others are more important to me than my accomplishments” (.23). Finally, deleting any of the items would not substantially alter reliability.
In addition, and preceding the hypotheses testing, scale means were computed and
descriptive statistics, skewness, kurtosis, and normality tests were conducted. The
interdependent self-construal appeared to be normally distributed, while ethical leadership, emotional exhaustion and follower’s voice were not. The skewness values were -.059, -.955, 1.052, and -.739 respectively, while kurtosis resulted to be -.160, 1.859, 1.927, and .440. The
29
Shapiro-Wilk test (Shapiro & Wilk, 1965) showed that the null hypothesis regarding the
normal distribution of the interdependent self-construal was not statistically significant (.72),
thus it cannot be rejected that the variable is normally distributed. For the other variables, the
test turned out significant and the hypotheses of normality could not be accepted. Similarly,
and supporting these results, the Kolmogorov-Smirnov test (KS test) demonstrates the
significance on the interdependent self-construal hypothesis (.20), while the other three
remain not significant. Although these outcomes may suggest some problems, the process
used by SPSS should reduce potential risks.
Results
Table 1 provides descriptive statistics and correlations of all variables studied. The
only significant correlations encountered are related to ethical leadership, which is significant
negatively correlated to emotional exhaustion (r=-.48; p=.01) and significant positively correlated to follower’s voice (r=.21; p=.05). These results, aligned with the proposed Hypothesis 1 and 2, imply direct relationships between the variables. However, no other
significant correlations have been found among the other variables.
Table 1
Means, standard deviations, correlation and reliabilities
Note: N=92
**. Correlation is significant at the 0.01 level (2-tailed). *. Correlation is significant at the 0.05 level (2-tailed).
Mean SD 1 2 3 4 1. Ethical Leadership 5.39 0.91 (.91) 2. Emotional Exhaustion 2.95 1.15 -.48** (.87) 3. Follower's Voice 5.44 0.68 .21* -.12 (.80) 4. Interdependent Self-Construal 5.05 0.58 .14 -.14 .19 (.78) Leader age 44.86 11.41 Employee age 36.21 12.08
Leader tenure with organization 16.50 13.73
Employee tenure with organization 11.78 12.09
30
Direct Effects
To further examine the direct relationships established on the correlation analysis, a
hierarchical regression was conducted to investigate the ability of ethical leadership in predicting levels of emotional exhaustion and follower’s voice, after controlling for employees’ tenure with leader. The dependent and control variable were entered in the first step, while the independent variable was inserted in the second step. The results of these
analyses are presented on Table 2 and Table 3.
Hypothesis 1 suggested that ethical leadership is negatively related to followers’
emotional exhaustion. In the first step, the control variable of tenure with leader was
introduced. This model was statistically significant F (1, 129) = 7.101; p=.009, and explained
the 5.2 % of variance in emotional exhaustion. In the second step, the variable of ethical
leadership was entered, and the total variance explained for the whole model raised to 23.2 %
(F (2,128) = 19.370); p=.000. The introduction of ethical leadership explained an additional
18 % of the variance in subordinates’ emotional exhaustion, after controlling for tenure with leader (ΔR²=.180; F (1,128) = 30.040; p=.000). In the final model, ethical leadership was statistically significant, with a Beta value of -.440 and p=.000. In other words, employees
who are managed by ethical leaders experience lower levels of exhaustion, and H1 is
supported.
Hypothesis 2 proposed that ethical leadership is positively related to followers’ voice.
As well as in the previous analysis,the control variable of tenure with leader was introduced
in the first step. In this case, this model was not statistically significant F (1, 129) = .187;
p=.666, and explained the 0.1 % of variance in followers’ voice. In the second step, the variable of ethical leadership was entered, and the total variance explained for the whole
model was 3.8 % (F (2,128) = 2.530); p=.084, still not significant. The introduction of ethical
31
with leader (ΔR²=.037; F (1,128) = 4.867; p=.029). In the final model, ethical leadership was
statistically significant (β = .198 and p=.029). Therefore, H2 is also supported, and the levels
of subordinates’ voice are higher for employees who are managed by ethical leaders.
Table 2
Hierarchical regression analyses for the relationship between Ethical leadership and Emotional Exhaustion.
Note: N=92
**. Correlation is significant at the 0.01 level (2-tailed). *. Correlation is significant at the 0.05 level (2-tailed). Control variable: employee tenure with leader.
Table 3
Hierarchical regression analyses for the relationship between Ethical leadership and Follower’s voice.
Note: N=92
**. Correlation is significant at the 0.01 level (2-tailed). *. Correlation is significant at the 0.05 level (2-tailed). Control variable: employee tenure with leader.
Moderation Effects
Hypothesis 3 and hypothesis 4 suggested a moderating effect of interdependent
self-construal on the relationships between the independent variable, ethical leadership, and the dependent variables, emotional exhaustion and follower’s voice. In order to test these interactions, a moderated regression analysis was used. However, before running the analysis,
R R² ΔR² B SE β t
Step 1 .228 .052**
Tenure with leader .080 .030 .228** 2.665
Step 2 .482 .232** .180**
Tenure with leader .039 .028 .111 1.383
Ethical leadership -.561 .102 -.440** -5.481
Emotional exhaustion
R R² ΔR² B SE β t
Step 1 .038 .001
Tenure with leader -.007 .017 -.038 -.432
Step 2 .195 .038* .037*
Tenure with leader .003 .017 .015 .165
Ethical leadership .140 .063 .198* 2.206
32
both independent variables (ethical leadership and interdependent self-construal) were
centered on their respective mean, by subtracting the mean from each score. Then, following
the method proposed by Aiken & West (1991), they were multiplied to create interaction
terms. In the Step 1 of the regression analysis, only the centered independent variables and
the control variable of tenure with leader were entered. The interaction between them was
added in Step 2. The only difference between the two analyses was in the consideration of the
outcome variable (Y), emotional exhaustion in the third hypothesis and voice in the fourth.
The results are presented on Table 3 and Table 4.
Hypothesis 3 stated that the negative relationship between ethical leadership and
followers’ emotional exhaustion is moderated by followers’ interdependent self-construal, so
that this relationship is stronger for higher levels of interdependent self-construal. The results
of the first analysis showed that this model was statistically significant F (3, 127) = 13.795;
p=.000, and explained the 24.6 % of variance in emotional exhaustion. In the second step, with the interaction entered, the total variance explained for the whole model raised to 28 %
(F (4,126) = 12.277); p=.000. The introduction of the interaction explained an additional 18 % of the variance in subordinates’ emotional exhaustion, after controlling for tenure with leader (ΔR²=.035; F (1,126) = 6.072; p=.015). In the final model, the interaction between ethical leadership and interdependent self-construal was statistically significant, with a Beta
value of -.208 and p=.015, meaning that negative relationship between ethical leadership and subordinates’ emotional exhaustion is stronger for higher levels of interdependent self-construal, and H3 is supported. A closer inspection of the conditional effects, and as it can be
extracted from Figure 2, for low levels of interdependent self-construal there is no effect on
the relationship. However, for high levels of interdependent self-construal, the levels of
33
Hypothesis 4 proposed that a positive relationship between ethical leadership and
followers’ voice is moderated by followers’ interdependent self-construal, so that this
relationship is stronger for higher levels of interdependent self-construal. Similar than in the
previous case, this model was statistically significant F (3, 127) = 3.062; p=.031, and
explained the 6.7 % of variance in followers’ voice. In the second step, the total variance
explained for the whole model was 12.9 % (F (4,126) = 4.656); p=.002. The introduction of
interaction explained an additional 6.8 % of the variance in voice, after controlling for tenure with leader (ΔR²=.061; F (1,126) = 8.871; p=.003). In the final model, the interaction between ethical leadership and interdependent self-construal was statistically significant (β =
-.277 and p=.003). However, and as showed in Figure 3, a more detailed examination
suggests that the effects of the interaction have opposite effects on individuals depending on
their levels of interdependent self-construal. For low levels of interdependent self-construal, followers’ voice increases with higher levels of ethical leadership. Contrary, individuals with high levels of interdependent self-construal experience a reduction of their voice as levels of
ethical leadership increase. Therefore, H4 cannot be supported, because the relationship is
34
Table 4
Moderation analysis of Interdependent self-construal on the relationship between Ethical leadership and Emotional exhaustion.
Note: N=92
**. Correlation is significant at the 0.01 level (2-tailed). *. Correlation is significant at the 0.05 level (2-tailed). Control variable: employee tenure with leader.
Fig. 2. Effects of ethical leadership and interdependent self-construal on emotional exhaustion.
R R² F ΔR² ΔF B SE β t
Step 1 .496 .246** 13.795
Tenure with leader .037 .028 .106 1.325
Ethical leadership -.539 .103 -.423** -5.237
Interdependent Self-Construal -.243 .162 -.118 -1.504
Step 2 .530 .280* 19.867 .035* 6.072
Tenure with leader .020 .028 .058 .714
Ethical leadership -.506 .102 -.397** -4.975
Interdependent Self-Construal -.111 .167 -.054 -.665
Interaction -.475 .193 -.208* -2.464
35
Table 5
Moderation analysis of Interdependent self-construal on the relationship between Ethical leadership and Follower’s voice.
Note: N=92
**. Correlation is significant at the 0.01 level (2-tailed). *. Correlation is significant at the 0.05 level (2-tailed). Control variable: employee tenure with leader.
Fig. 3. Effects of ethical leadership and interdependent self-construal on followers’ voice.
R R² F ΔR² ΔF B SE β t
Step 1 .260 .067* 3.062
Tenure with leader .004 .017 .022 .251
Ethical leadership .122 .063 .173 1.923
Interdependent Self-Construal .199 .099 .174 2.002*
Step 2 .359 .129** 11.933 .061 8.871
Tenure with leader -.008 .017 -.042 -.472
Ethical leadership .146 .062 .207 2.352*
Interdependent Self-Construal .296 .102 .259 2.907**
Interaction -.349 .117 -.277 -2.978**
36
Indirect Effects
Hypothesis 5 suggested that the positive relationship between ethical leadership and
subordinate’s exhaustion is mediated by followers’ voice. In order to test that proposition, a multivariate analysis of four regressions was used. The results can be found on Table 5. As in
the previous analysis, the control variable of tenure with leader was considered. In the first
step, the independent variable, ethical leadership (X), and the outcome, emotional exhaustion (Y) were introduced. The β coefficient (-.440) was statistically significant with a p=.000. In the second step, the simple regression was calculated with the independent variable, ethical leadership (X), and the mediator, followers’ voice (M). In this case, the β=.198 was also significant (p=.029). In the last two steps, followers’ voice and ethical leadership were both considered, and the last coefficient was higher than the first one (β=-.455) and significant as well (p=.000). According to the statement proposed, the beta should decrease when voice is
included in the analysis. However, the change expected is not satisfied and it suggests that
voice does not mediate the relationship between ethical leadership and emotional exhaustion.
In other words, Hypothesis 5 has to be rejected.
Table 6
Mediation analysis of Followers’ voice on the relationship between Ethical leadership and Emotional exhaustion.
Note: N=92
**. Correlation is significant at the 0.01 level (2-tailed). *. Correlation is significant at the 0.05 level (2-tailed). Control variable: employee tenure with leader.
B SE β t B SE β t
Step 1 Ethical Leadership (X) -.561 .102 -.440** -5.481
Step 2 Ethical Leadership (X) -- -- -- -- .140 .063 .198* 2.206
Step 3 Followers' Voice (M) .040 .143 .022 .279
Step 4 Ethical Leadership (X) -.566 .105 -.455** -5.414