Amsterdam Business School
Influence of Leader-member exchange on employee perceived control over performing well with moderating effect of the use of PIs for incentive purpose
Name: Olena Kotyuk Student number:11098015
Thesis supervisor: Sjors van der Heide Date: 20 June 2016
Word count: 11.239
MSc Accountancy & Control, specialization Control
Statement of Originality
This document is written by student Olena Kotyuk 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.
Abstract
The present study presents an analysis aimed to explain and to predict behavioural changes in perceived control of employee caused by quality of leader –member exchange.
Using the theory of LMX and theory of planned behaviour I try to explain how the high quality exchange relationship between manager and employee in a dyadic dimension could positively influence the perceived control of the employeeto perform work related tasks.
The LMX theory states the high-quality leader–member relationships or exchanges (LMX) are characterized by high levels of trust, interaction, support, and formal and informal rewards (Dienesch & Liden, 1986). Behavioural theory argues that perceived control of the employee will increase in a trustworthy and informal work environment (Fishbein & Ajzen, 1975). I also attempt to analyse whether there is a moderation effect of the level of use of the performance indicator for incentive purposes on the relationship between the quality of exchange and perceived control of the employee. The choice of this moderator is argued by the fact that, according to several academic studies, personalbeliefs of a person's ownabilities to perform could be altered in the moment the person knows he is being evaluated and rewarded according to the performance standards (Studies conducted in the 1920s at the Hawthorne Plant of the Western Electric Company (Anthony A. Atkinson, Robert S. Kaplan, Ella Mae Matsumura, S. Mark Young, Management accounting, 6th Edition, 2012)).
We surveyed 106 pairs of operational employees and their immediate managers in various jobs and industries and tested our hypotheses with the set of multiple linear regressions.
The quality of the LMX was measured by two separate constructs: manager’s perception LMX and employee’s perception LMX.
Results showed that from the manager’s perspective of LMX the quality of the exchange relationship positively influences the perceived control of the employee to perform well; however, there is no evidence of any association between the quality of dyadic relationship and perceived control from the perspective of employee perception of LMX. We did not find any interaction effect of the use of performance indicators for incentive purposes on the relationship between the quality of exchange and perceived control. These results give a direction for future research. Overall, this study demonstrates that quality of exchange relationship between manager and employee can produce a positive influence on the behavioral antecedents of performance such as perceived control to perform well. We discuss implication for managers who want to initiate development of high-quality exchange relationship with their employeesand discuss the way to implement them.
Keywords: leader-member exchange, perceived control over performing well, performance indicators for incentive purposes, theory of planned behavior, LMX theory.
Table of contents
1. Introduction ... 5
2. Research area ... 6
2.1 Research question and backgrounds ... 6
2.2 Motivation of study ... 8
3. Literature reviews and hypothesis ... 9
3.1 Leader-Member Exchange (LMX) ... 9
3.2. Measurement of LMX ... 9
3.3 A Theory of Planned Behaviour ... 10
3.4 Perceived control over performing well ... 11
3.5 Hypothesis development 1 ... 13
3.6 Performance indicators for incentive purpose ... 14
3.6.1 Description of performance indicator (PIs) ... 14
3.6.2 A theory of performance measurement properties ... 15
3.7 Hypothesis development 2 ... 16
4. Research methodology ... 18
4.1 Respondents ... 18
4.2 Survey Instrument ... 19
5. Research findings ... 20
6. Conclusion and limitation of the study ... 31
6.1 Discussion ... 31
6.2 Limitations ... 34
6.3 Conclusion ... 35
References ... 36
1. Introduction
The search for and identification of those qualities that increase a leader's effectiveness has been a major concern of practicing managers and leadership researchers for the past several decades (cf. Bass, 1981; House, 1971; 1988; House & Baetz, 1979; Stogdill, 1974; Yukl, 1989a; 1989b). One may think that a good leader has an enormous impact on the development of the organization and performance of employees. The correct strategy, plan of action and ability to motivate and guide team members are decisive characteristics of a leader when pursuing organizational growth. However, the effectiveness of a leader cannot be evaluated per se; it always establish itself through the interconnection of the leader with other members of the team (Graen & Uhl-Bien, 1995; Liden, Sparrowe, & Wayne, 1997). These interconnections identify themselves in the concept of leader-member exchange (LMX). The most relevant contribution of the leader-member exchange approach to the research dominion of leadership is that leaders can develop a high-quality exchange relationship with their subordinates (RT Sparrowe, RC Liden - Academy of management Review, 1997). The quality of the member’s exchange relationship with the leader is based on the degree of emotional support and exchange of valued resources, and is crucial in determining the members’ perception regarding their ability to perform the relevant work tasks (Liden, Sparrowe & Wayne, in press). Low level of perceived behavioural control of individuals are characterized by an understated degree of self-perception about one’s own abilities to perform effectively. According to academic research (Diener & Dweck, 1978), this soon devolves into worry, doubts about one’s capacities, and irrelevant digressions. On the other hand, employees who are more confident and optimistic are more likely to select challenging tasks, set high goals, deal productively with obstacles and setbacks and retain access to advanced problem-solving abilities.
One of the commonly used factors of influencing employees’ behaviour and their attitude towards the performing of tasks is the use of performance indicators (PIs) aimed to evaluate and incentivize employees’ performance (Gibbs et al. 2004).
It could be the case that the employment of performance indicators as instruments for evaluating and rewarding employees can foster the development of the exchange relationship. In turn, the high-quality exchange relationship between manager and subordinates will positively affect perceived behavioural control of the employee over his ability to work well. The proving of this statement could provide a valuable contribution to the research about
domain.
In context of my current research, the following definitions apply:
Leader-membership exchange
The extent to which an employee feels that he has a quality relationship with his manager (Gómez & Rozen, 2001).
Perceived control over performing well
The extent to which employees believe to be capable of always meeting all job requirements ( Fishbein & Ajzen, 2010)
Use of PIs for incentive purpose
Metrics used to quantify the efficiency and effectiveness of actions for evaluation and rewarding purposes (Neely, Gregory, & Platts, 995).
2. Research area
2.1 Research question and backgrounds
The main goal of this study is to examine how the effect of the quality of dyadic relationships between a supervisor and his subordinates influence the perceived control of employees to perform well, and how this relationship changes in the presence of performance indicators for incentive purposes. Perceived control over performing well is one of the important determinants of performance (B.A.C. Groen, C.P.M. Wilderom, and J.F.Wouters). According to the Fishbein, M & Ajzen, I (2010) people form beliefs about personal and external factors that can encourage or hinder their attempts to perform the behaviour. Indeed B. Groen in her research states that perceived control over performing well constitutes one of the main behavioural factors that influences the performance of employees. A. Bandura emphasizes how cognitive, behavioural, personal and environmental factors interact to determine motivation and behaviour (Crothers, Hughes, & Morine, 2008).Since an organization’s performance is a function of the aggregate behaviour of its members, organizational improvement can occur only when members’ behaviour change (Goodman and Dean, 1982; Tannenbaum, 1971). In this research I am going to investigate more precisely the effect that high-quality leader-member exchange (LMX) relationships produce on the behavioural control of employees. However, the organizational environment is never simple and straightforward; there are other factors, for example the implementation of a new ERP system, balanced scorecard system or other performance evaluation system that can significantly undermine the self-belief of
employees to perform well. As I am analysing behavioural responsiveness of employees to the quality of the dyadic relationship between them and their direct supervisors it would be appropriate to investigate how this relationship could be influenced by significant external factors. Since performance evaluation usually affects employee compensation and rewards (Kaplan and Atkinson, 1998; Horngren et al., 2002) and employees are likely to be concerned with the performance evaluation process, I decided to investigate how the use of performance indicators (PIs) for incentive purposes can influence the LMX relationship and employees’ perceived control over performing well.
According to B. Groen’s research, the performance indicators for incentive purposes do not always produce a positive effect on the ability of employees to perform well, due to the excessive stress experienced by them. The perception of an employee of the evaluation instruments used for incentive purposes could influence the employees' control beliefs regarding their ability to perform the task. Indeed, considerable management accounting research was done on the behavioural consequences of performance evaluation methods and procedures (Klein, 1991; Mowen, Middlemist, & Luther, 1981; Wright, 1989). Also several studies (e.g. Hopwood, 1972; Otley, 1978; Ross, 1994; Lau and Buckland, 2001) have found a significant correlation between performance evaluation and quality of relationship between employee and supervisor. By utilizing LMX literature, theory of planned behaviour and performance measurement instruments I will try to offer a new insight into the area of behavioural studies by examining the perception of individuals of their abilities to perform well. Specifically, I am going to investigate how the use of performance indicators can enhance the employee’s perceived control over performing well in the high-quality exchange relationship and diminished negative effect of poor quality exchanged relationship.
2.2 Motivation of study
This study contributes to both scientific and societal points of view. From the scientific point of view, the questions I am going to investigate will expand the existing literature regarding the relationship between the LMX and perceived control of employees over performing well. During the three past decades, there has already been some research regarding the influence of the quality of dyadic employee-supervisor relationship on different outcomes of the organization. However, there are many more possibilities to study this relationship throughout moderating effect of a variety of organizational factors that could produce different and often unexpected impact on the aforementioned relationship (Ahearne, M., Mathieu, J., & Rapp, A.,2005). In choosing the moderating variable I was driven by the idea to find the variable that will be common and actual for many companies and which has a widespread employment in organizational management. Following this reasoning, I decided to investigate the impact of the use of performance measurement indicators for incentive purpose as the moderating variable. There were not enough studies on this topic in previous literature, therefore, answering this research question might provide the existing literature more “power” or just decrease the credibility of past research or give new findings.
From a societal point of view, the results of the investigation shed light on how the introduction of the performance measurement system for incentive purpose affect employees'perception of their abilities. This knowledge, in turn, give managers more insight into what motivate employees to perform well. Knowing this can be crucial for the effective distribution of limited resources within the organization, designing the management control system and developing the actions for achieving performance targets.
3. Literature reviews and hypothesis
3.1 Leader-Member Exchange (LMX)
LMX theory has its roots in vertical dyad linkage theory (VDL) developed by Graen and his colleagues Cashman et al. 1976, Dansereau et al. 1975, Graen 1976 and Graen & Cashman 1975. Leader-member exchange (LMX) is operationally defined as the extent to which an employee feels that he has a quality relationship with his manager (Gómez & Rozen, 2001). The basic premise of this theory was that leaders differentiate between subordinates in the way they approach them (Graen & Uhl-Bien, 1995). Indeed, hierarchical relationships vary in terms of support and openness, which Crouch and Yetton (1988) define as trustworthiness. A leader develops a much closer relationship with some subordinates (in-group) and gives them more room for negotiating than other subordinates (out-group) Cashman et al. 1976 and Dansereau et al. 1975.
The relationship is initiated through a series of interpersonal exchanges. Leaders and followers in these high-LMX relationships often demonstrate enhanced levels of satisfaction and effectiveness, as well as mutual influence, more open and honest communication, better access to resources, and more extra-role behaviours (Gerstner & Day, 1997). Low-quality relationships, in contrast, tend to disadvantage employees and affect their control beliefs (Vecchio, 1997). In low-quality relationships, members receive less access to the supervisor, fewer resources, and more restricted information, potentially leading to dissatisfaction in the job, lower organizational commitment and lower self-efficacy (Gerstner & Day, 1997). In the last three decades, however, the LMX theory underwent significant evolution, discarding the in-group and out-group labels. In the current version of the theory the quality of relationship is measured on a continuum. Thus, LMX is widely used to describe the quality of interaction between manager and subordinate ( Dienesch & Liden 1986 and Uhl-Bien, Graen, & Scandura 1997).
3.2. Measurement of LMX
There are two major questions regarding the LMX approach derived from literature and research. These are questions of dimensionality and content of exchange. According to LMX
efforts toward the job or favorable task assignments were at heart of the research. However, in a recent review of LMX literature, Liden et al. (1997) highlight that LMX is not based solely on the job-related elements but may also include socially related features. According to this study, Dienesch and Liden (1986) and Liden and Maslyn (1998) developed four dimensions of LMX relationships named contribution, affect, loyalty and mutual trust, and professional respect.
Contribution
According to Settoon, Bennet and Liden (1996) a high-quality relationship leader will provide beyond employment contract assistance and the followers will execute the leader’s beyond employment contract requirements (p. 220). Leaders can be sure that subordinates will act as loyal followers. The subordinates, in turn, can be sure that leaders will protect and support them (Graen & Scandura, 1987; Graen & Uhl-Bien, 1995; Uhl-Bien et al., 2000; Wayne, Shore, & Liden, 1997)
Loyalty and mutual trust
According to Uhl-Bien et al. (2000) trust exists when a party in a LMX relationship believes that the other party will not betray nor manipulate his vulnerability, believes that he can count on the other party and believes that he and the other party share common values when acting toward each other.
Affect
Affect reflects the friendship and liking that the dyad member feels toward the other and reflects in terms of exchange of “favours” among of them (Uhl-Bien et al., 200, p. 161). That is, in the high-quality relationship the members exchange their favours freely without being ask and without expectation of payback (ibid, 2000, p. 161)
Professional respect
Professional respect is measured in terms of knowledge and competence that parties possess. A leader respects a member who is highly capable and who will not damage his or her reputation. A follower respects a leader in relation to his diligence and professional disposition.
3.3 A Theory of Planned Behaviour
behaviour is defined by their attitude, perceived norm and perceived control to perform the behaviour. To explain the positioning of the concepts of perceived control over performing well I report two possible versions of the theory of planned behaviour developed by Ajzen and Madden.
Figure 3.1 Ajzen and Madden, Theory of planned behaviour.
The first model assumes that perceived behavioural control has motivational implications for intentions. It means the the perceived control produce an indirect effect on behaviour and is completely mediated by intention. On the other hand, the second version is based on the assumption that perceived behavioural control produces a direct influence on behaviour. According to Ajzen and Madden the theory of planned behaviour states that execution of action will depend not only on motivation but also on adequate control over the action in question. It implies that perceived behavioural control can predict execution of behaviour independent of behavioural intention. In other words, according to the second version of the theory of planned behaviour perceived behavioural control can also act as a partial substitute to the actual control that makes the action happens. This implies that perceived control can predict the behaviour and thus be considered as determinants of human actions (I.Ajzen, 1991).
3.4 Perceived control over performing well
Perceived control is belief about resources, possibilities and other factors that could facilitate or impede the behavioural performance. A direct measure of perceived behavioural control should capture people’s confidence that they are capable of performing the behaviour under investigation (Bandura, 1977,1997)
A number of different items have been used for this purpose. Some deal with the difficulty of performing the behaviour, or with the possibility that the person could do it. Such items are often intended to capture the respondent’s sense of self-efficacy with respect to performing the tasks. Other items used to assess perceived behavioural control refer to the behaviour’s controllability. These items address people’s beliefs that they have control over the behaviour, and that they can meet all job requirements.
Essentially, control beliefs could be of internal and external orientation. Those of internal orientation refers to the extent that individuals believe that events in their lives are under their own control. Conversely, external orientation is determined by forces from outside themselves, such as luck, fate or chance. The prevailing view of the relationship between LMX and control holds that the quality of dyadic relationships among the organization's members influence and define the level of the beliefs of employees’ capabilities to perform behaviours at planned levels (E.A. Skinner, J.G. Wellborn, J.P. Connell, 1990)
Researchers are becoming increasingly interested in the role of control over desired behaviour. The theory of reasoned action states that the stronger a person’s intention, the more the person is expected to try, which will result in a greater likelihood that the behaviour will actually be performed. However, the intention as the sole predictor of behaviour will be insufficient whenever control over the behaviour is incomplete (D. J. Terry and J. E. O'Leary, 2011). That is, limited control over behaviour make it difficult to predict the expected behaviour. Ajzen (1985; Schifter & Ajzen, 1985) in his theory of planned behaviour proposes that the control beliefs of the individual ultimately determine his actions, and latitude of these beliefs is directly related to the presence or absence of resources and opportunities. The more resources and opportunities individuals think they possess, and the fewer obstacles or impediments they anticipate, the greater should be their perceived control over the behaviour (Fishbein & Ajzen, 1975). That means that the level of available resources defines the level of employee perceived control. Within a working environment, the availability of resources can be ensured by effective sharing of information, reciprocal help, common discussions on ongoing tasks and interchange of working experiences.
Developing of a trustworthy environment can open a free flow of information and interaction between employees and their managers, contributing to increasing the personal confidence of employees and lowering of working stress (P.E. Spector, 1986).
3.5 Hypothesis development 1
Based on the theory and prior research set out in the previous paragraphs, below I will present my first hypothesis.
In this study, I use the concepts of psychological theories such as theory of planned behaviour and LMX theory to explain the effects of the leader-membership exchange on the perceived control over performing well. According to these theories, psychological factors play a crucial role in explaining the behaviour of employees. I assume that when the employee and his supervisor have positive interaction with each other, which is characterized by trust, professional respect, reciprocal availability and support, the perception of behavioural control of the employee will increase.
According to LMX theory, the supervisor shares personal and professional resources (information, assistance in decision making, task assignment, support and attention) in exchange for employee performance on relevant tasks (Graen & Cashman, 1975). Mainly the interaction between employee and supervisor has double effects. From the part of the supervisor, it means provision of the employee with all necessary information to enable him to perform in the way that leads to achievement of organizational objectives and goals (high-quality relationship). From the part of employees, they need continuous exchange of information in order to be able to have access to the necessary resources and to receive the feedback about the work related tasks (perceived control over performing well) (Ajzen,2012: Fishbein &Ajzen, 2010: Yzer, 2012). In other words, when the relationship between the supervisor and subordinates are based on openness, trust, professional respect and support it creates a flourishing environment for employees to work well. In such high-quality relationship employees will perceive more control over the external factors and more power over overcoming obstacles and impediments in working processes (E.A. Skinner, J.G. Wellborn, J.P. Connell, 1990)
The above leads to my first hypothesis:
3.6 Performance indicators for incentive purpose
3.6.1 Description of performance indicator (PIs)
In recent times, interest in performance measures has grown significantly as evidenced by the large literature on ‘benchmarking', ‘total-quality' measures and ‘balanced-scorecards' (Borthick and Roth, 1997, Cooper, 1989). Many organizations have started the introduction of these indicators, which include both non-financial as well as financial measures. The increased focus on performance evaluation by managers and academics reflects the increased pressure that organizations experience to improve performance. Moreover, cheaper information technology allows various measurements to be made relatively easily (Foster, 1996, Johnson, 1995, Johnson and Kaplan, 1987 and Lynch and Cross, 1991).
A performance measurement system plays numerous roles, they include (Franco Santos et al., 2007: Bourne and Bourne, 2007):
• Communication direction – what is have to achieved • Influencing behaviour – in positive or negative way • Stimulating action – identifying moment of intervention • Facilitating learning – both single and double loop
• Implementation of strategy – ensuring organizational objectives are achieved
In the wider sense, performance indicators (PIs) are “everything used to measure the job performance of employees”. Concretely, performance indicators are measures used to quantify the efficiency and effectiveness of actions (Neely, Gregory, & Platts, 1995). Examples are client satisfaction, process efficiency, timeliness, and quality of work completed.
In the purpose of this research, we will look at the performance indicators from the perspective of their incentive functions. Specifically, use of PIs reflects the extent to which managers use the metrics for monetary compensation, non-monetary rewards, and evaluation purposes. I use the term of PIs in its complexity as a combination of monetary, non-monetary and evaluation performance indicators. It may include salary increase, bonuses, or extras, promotion or authority within the organization, and official performance ratings. From the literature, we know that one of the purposes of incentive instruments in a performance measurement system is to align the motives and actions of the agents and principals (Jenkins Jr., Mitra, Gupta &
Shaw, 1998; Stajkovic & Luthans, 1997,2003). However, incentive purpose performance indicators have other functions that play an important role in defining the intensity and quality of the communication and interaction between employees and supervisor, such as facilitating decision-making and motivating and controlling of employees (Demski&Feltham, 1976: Sprinkle, 2003)
Following the preceding reasoning we can suggest that performance indicators accomplish the function of the "bridge" in the process of interaction and information sharing between employee and supervisor. First, they provide the guidelines and motivation for employees, and second they foster the communication process between the principals and agents assuring in such a way achievement of organizational objectives.
3.6.2 A theory of performance measurement properties
In order to examine the effect that performance indicators for incentive purposes might have on the relationship between quality of the dyadic relation and perceived control over performing well, it is appropriate to look at the properties of the performance measures. A central property of a performance measure is its risk profile and it includes two types of risk: uncontrollable risk and controllable risk. The uncontrollable risk represent noise originated by all environmental factors that affect the performance measure and that the employee is unable to control. Conversely, controllable risks reflect environmental uncertainty affecting the performance measure that employee can control or react to (Gibbs et al. 2009). Others have called controllable risk "specific knowledge" (Jensen and Meckling 1992, Raith 2008). It implies that the employee has knowledge that is valuable to the firm, and performance indicators for incentive purposes can be used to motivate the employee to exploit this knowledge (Shi 2011). The uncontrollable factor, instead, can materially weaken the perception over being able to complete the tasks (Gibbs et al. 2009, Hall, 2008).
Another important property of performance measure is distortion, which appears when the employees’ actions have different impact on the performance measures than expected (Holmstrom and Milgrom 1991, Baker 1992, Feltham and Xie 1994, Baker 2002).
The perfect design for performance indicators includes minimal risk and consistency with organizational goals. However, it is usually the case that the undistorted measures contain a lot of uncontrollable risks and vice versa (Baker 2002). One common way of improving the
et al. 2004)is use of the subjective evaluation. However, subjective evaluations are found to be positively related to LMX (DelVecchio, 1998). There is also the research evidence suggesting that managers tend to evaluate high LMX employees higher than their objective performance rating (Duarte, Goodson, &Klich, 1994). Nevertheless, some studies fail to confirm this relationship (DelVecchio, 1998).
3.7 Hypothesis development 2
Thus far we have emphasized the importance of the interaction, communication and information sharing between the employee and supervisor for planning, resource allocation, decision-making, acting, monitoring and improving the working process. However, the different studies in management accounting fields proof that with the introduction of performance measures in the performance evaluating process the employees' behaviour tend to alter when they know they are being studied and their performance is being measured (Study conducted in the 1920s at the Hawthorne Plant of the Western Electric Company (Anthony A. Atkinson, Robert S. Kaplan, Ella Mae Matsumura, S. Mark Young, Management accounting, 6th Edition, 2012)). As outlined in the previous paragraph the important factors that drive employee intention to perform are the existence of the control over the behavioural goal and existence of the performance measurement system which creates the incentive to perform the assigned task. The presence of the performance indicators for the incentive purposes provide employees with additional motion, increasing their intention and motivation to complete the task. Indeed, control systems often consist of performance measures and reward instruments (Banker&Datar, Jensen & Meckling,1992). Performance measures point out the direction to follow for employees, and the compensation system gives the incentives to take this direction. Following reasoning of the previous chapters we can assume that with the appearance of the performance measurement system the demand in communication and reciprocal interaction between the supervisor and employees will increase. The employees feel a natural need to assure the achievement of the target imposed by performance measures as they know that they will be evaluated in relation to this target. Any factor outside the control or competence will stimulate him to deepen the communication with his manager. Therefore, employees in order to diminish the uncontrollable risks of performance measures would be likely to develop a positive relationship with their supervisor and collaborate together towards achievement of the targets. That said the presence of a performance measurement system increase the
supervisor-employee relationships and perceived control of supervisor-employee to perform well. I predict my second hypothesis to be:
H2: LMX together with use of the performance indicators for incentive purposes will lead to an increased level of perceived control over performing well.
4. Research methodology
Data were collected from employees and their immediate managers by employing two online surveys: one among managers and another for one of their subordinates. Following section contains details about the respondents and survey instruments.
4.1 Respondents
The survey was completed by 106 pairs of respondents. The respondents represent the dyadic pairs of employees and his direct managers who have met the requirements of the survey: (1) the non-managerial employees had to be professionals or members of staff who carry out the work; (2) they must have worked in their current function for at least one year; (3) the supervisory managers had to use the performance measure to measure their employees’ performance. The difficulty to find the dyadic pairs of respondents was mitigated by expanding the network among the students who later could benefit from the common created database. The confidential treatment was assured for every respondent. For every person who participated in the survey was sent the research report showing his or her personal scores. The online survey was internationally based, so the data was collected also from outside The Netherlands. Other demographic statistics are provided in Table 3.1.
Table 4. 1 Respondents characteristics
Characteristics Employees (N=104) Managers (N=99)
Gender ª % Male 50 64
% Female 48 29
Education ᵇ % Lower vocational education 2
% Primary education 1
% Intermediate general education 2 3
% Intermediate vocational education 11 7
% Higher general education 10 6
% Higher vocational education 26 23
% Scientific education 47 61
Age Mean (SD) 33 (9.6) 39 (9.6)
Tenure Mean (SD) 6 ( 5.9) 7 (6.8)
From the 106 pairs only 99 managers and 104 employees gave the responses regarding their personal data (gender, education, tenure). Employees are almost equally represented by the male and female, 50% and 48% respectively (2% of employee and 7% of manager did not specify their sex). The manager males are almost as twice as much as the manager women, 64,2 % and 29,2% respectively. The average tenure in the current position are 6 years (SD 5.9 years) for employee and 7 years (SD 6.8 years) for managers. Interesting fact is that 84% of employee holding higher education and 47 % of employee holding scientific education. We can conclude that majority of the employees of our sample are highly skilled and possess the high level of knowledge.
4.2 Survey Instrument
The survey questionnaire was composed in Dutch for Dutch speaking participants and in English for English-speaking participants. All the items had a 7-point fully anchored Likert scale: (1) totally disagree, (2) disagree, (3) moderately disagree, (4) neutral, (5) moderately agree, (6) agree, (7) totally agree. An overview of the items is given in the appendix. Individual differences in demographic factors can have an impact on the relationship between the variables. Therefore, demographic data like age, gender, educational level, employee and manager tenure were collected from all the respondents.
LMX was assessed by selecting 7 question from two scales: the LMX-7 (Graen & Uhl-Bien, 1995) and the LMX-MDM developed by Liden and Maslyn (1998). The content of both manager and employee questionnaire were the same. According to Dienesch and Liden (1986) and Liden and Maslyn (1998) the quality of the exchange is estimated based on the following characteristics: contribution, loyalty, affect and professional respect. Other LMX research has produced measures of these constructs and demonstrated validity of these dimensions (Liden & Maslyn, 1998; Schriesheim, Neider, Scandura, & Tepper, 1992). The questions:” I understand his/her problems and needs in his/her work”, “I recognize his/her potential”,” I would “bail him/her out” at my expense” aimed to reflect the quality of relationship characterized by the leader’s flexibility, supportiveness, use of feedback, likability, loyalty, and trustworthiness (Ferris et al., 2009; Graen & Scandura, 1987), as well as the generic effectiveness of the leader–member relationship itself (Graen, Liden, & Hoel, 1982).
Perceived control over performing well were developed based on standard guidelines for constructing theory-of-planned-behaviour questionnaires (Darker & French, 2009; Fishbein &
survey were chosen the best fitting 3 items. The constructs of perceived control over performing well was compute by collecting the data from the employee’ questionnaire. I measured the PIs for incentive purpose selecting the items that reflect monetary, non-monetary incentives and PIs for evaluation purposes used by managers to evaluate and reward their employee. They include financial incentives (bonuses, salaries), non-financial incentives (promotion, increase in rating, reputation factor) and PIs used for evaluation purposes. According to the Grafton (Grafton et al., 2010) all three types of incentives usually used in tandem with each other. For instance, the same performance indicators can be used for evaluating scope as well as for rewarding employees. The construct of PIs for incentive purpose was measured by scales on performance metrics (Moers, 2006) that was collected from managers’ questionnaire.
All questions used to measure each variable are illustrated in the Appendix.
As control variables I used: employee gender, educational level, age, and professional tenure. These demographic variables may provide an alternative explanation for the differences in the ratings of perceived control (T. A. Scandura, G.B. Graen. 1984).
5. Research findings
Since the reliability analysis revealed low level for one item belonging to perceived control over performing well (“Certain condition make it impossible for me to always meet everything that is expected of me in my work.”) I have excluded it from the model. The initial alpha coefficient was 0.309, the analysis showed that by deleting this items the alpha coefficient could be improved to an acceptable level of 0.708. The reason for such a low coefficient alpha was due to the negative formulation of the question that in turn lead to the distortion of the reliability coefficient. Similar outcomes had been found in other research on the theory of planned behaviour (Yzer, 2012). Therefore, consistent with most theory-of-planned-behaviour studies (Bleakley & Hennessy, 2012; Fishbein & Ajzen, 2010), we only used the positively formulated item to measure perceived control over performing well. All other items were retained as they did not present any negative impact on the reliability of the constructs of perceived control, management and employee LMX and PIs for incentive purposes.
The quality of LMX were measured by two separate constructs that represent two perspectives of quality of LMX from managers’ point of view and from employees’ point of view. According to the Gerstner and Day’s (1997) meta-analysis of the LMX literature in which they reported low correlation, r 1⁄4 .29 (r corrected 1⁄4 .37), between these two concepts, leader LMX and Member LMX are two different constructs.
In order to detect items retention, I used a principal component analysis with varimax rotation. The results of factor loadings and descriptive statistics of the items, together with reliability coefficient are shown below:
Table 5.1 Descriptive statistics and factor loading s of the measurement model
α N M SD Min Max SD
Percieved control to perform well 0.708
Item 1 106 5.1 1.33 1 7 0.78 Item 2 106 4.58 1.45 1 7 0.87 Managers' purspective of LMX 0.845 Item 1 103 5.91 0.81 2 7 0.76 Item 2 103 5.91 0.91 2 7 0.81 Item 3 103 5.98 0.98 1 7 0.78 Item 4 103 5.82 0.92 2 7 0.73 Item 6 103 5.62 0.98 2 7 0.69 Item 7 103 6.27 0.91 1 7 0.82
Use of Pis for incentive purpose 0.887
Item 1 103 5.47 1.23 1 7 0.60 Item 2 103 5.57 1.33 1 7 0.72 Item 3 103 5.45 1.39 1 7 0.82 Item 4 103 5.29 1.62 1 7 0.86 Item 5 103 5.41 1.49 1 7 0.87 Item 6 103 5.26 1.44 1 7 0.77 Employees' purspective of LMX 0.908 Item 1 106 5.7 1.17 2 7 0.76 Item 2 106 5.59 1.20 2 7 0.81 Item 3 106 5.6 1.11 2 7 0.78 Item 4 106 5.7 1.11 1 7 0.88 Item 5 106 4.96 1.37 1 7 0.78 Item 6 106 5.45 1.24 1 7 0.80 Item 7 106 6.01 1.00 1 7 0.82 Variables
As I rely on self-reported measurement, I retained only strong factors loading 0.5 or higher on the target construct. As one of the items of the management LMX construct (“I would “bail
rotated solution in Table 5.1 shows four factors corresponding to four constructs. All the items loaded highest on their own construct.
Table 5.2 provides additional descriptive statistics, n, means, and standard deviation for all the scales totals.
Table 5.2 Summary statistics of the constructs of measured model
Variables N Minimum Maximum Mean Std. Deviation
PC 106 1 7 4.84 1.23
MTLMX 103 2 7 5.87 0.71
ETLMX 106 2 7 5.57 0.94
PIs 103 1 7 5.41 1.14
PC – Perceived control to perform well, MTLMX – Managers total LMX, ETMLX – Employee total MLX, PIS – Use performance indicators for incentive purpose; Valid N (listwise) – 103 dyads.
From the descriptive statistics we can notice that both self-reported data from managers and employee perspectives report relatively high quality of their relationship that are above the scale midpoint, however, the mean of the MTLMX is higher than the mean of the ETLMX, 5.87 and 5.57 respectively. This explains that managers perceive higher level of quality of their relationship with their employees then employees do. As managers are often evaluated on their leadership skills, which in turn depend on the quality of the relationship with their employees (Leslie, Jean Brittain; Fleenor, John W.,1998) it may be the case that supervisors tend to think that they are good to create positive connection with their subordinates and overestimate the quality of the exchange relationship.
Table 5.3 provides the results of the correlations coefficient of both Pearson and Spearman tests. Table 5.3 Bivariate correlation matrix
1 2 3 4 5 6 7 8 0.011 0.027 0.062 .267** 0.099 0.055 0.119 . 0.915 0.783 0.546 0.006 0.323 0.577 0.236 0 -0.063 -0.131 0.124 0.042 0.008 0.173 0.998 0.524 0.197 0.204 0.672 0.935 0.081 0.006 -0.158 .643** -0.029 -0.04 -0.092 0.002 0.952 0.107 0 0.767 0.692 0.352 0.986 0.114 -0.185 .630** 0.025 -0.05 -0.085 0.069 0.268 0.067 0 0.803 0.628 0.405 0.504 .212* 0.087 -0.091 -0.033 .206* 0.155 .242* 0.031 0.374 0.354 0.746 0.037 0.113 0.014 0.176 0.1 -0.108 -0.155 .234* 0.122 .295** 0.078 0.314 0.28 0.133 0.017 0.219 0.002 0 0.036 -0.041 -0.153 .226* 0.141 0.021 0.999 0.714 0.677 0.131 0.02 0.157 0.83 0.113 0.169 -0.093 0.044 .254** .256** 0.086 0.259 0.088 0.352 0.672 0.01 0.009 0.387 Managers total LMX Employees total LMX
PIS for incentive purpose 7 8 Pearson correlation Items 1 2 3 4 5 6 Spearman's rho Employee gender ª Employee education ᵇ Employee age Employee tenure PC
Notes: n=106; **p<0.01; *p<0.05; Pearson value are below diagonal ª 1-male, 2- female
ᵇ 1 - Lower vocational education, 2- primary education, 3 – Intermediate general education, 4 – Intermediate vocational education, 5 – Higher general education, 6 – Higher vocational education, 7 – Scientific education.
The results of correlation show that the quality of the dyadic employee-supervisor relationship appears to produce the positive association with the perceived control over performing well only from the perspective of managers (0.234, p<0.05). From the perspective of employee there is no significant relationship between the quality of the exchange employee-supervisor and perceived control over performing well. It could be explained by the fact that employees perceive their ability to perform a task independently from their perception of the quality of the relationship. However, we observe different situation when take to account the managers’ point of view of LMX. The possible explanation of these phenomenon may be that supervisor and subordinates searching for the different qualities of exchange from their partners. According to Dockery and Steiner (1990) and Day and Crain (1992) employees look for social and manager look for work-related currencies.
We can notice that perceived control over performing well is highly correlated with the use of performance indicator for incentive purpose in the evaluation and rewarding of employees (b=0.254, p<0.01) This fact indicates that perception of employee of their ability to perform the task could be influenced by the level of use of the performance indicators which managers use for rewarding the employees. As was already found in previous literature incentives are the effective motivator for employee
(
Rynes et al, 2004).We can also observe the high level of correlation between the managers’ perception of the quality of the relationship with their employees and use of the PIs for incentive purpose (b=0.256, p<0.01). It may imply that managers tend to highly evaluate and reward those employees with whom they have high quality relationship (Friel, 2004).
There is no significant relationship between the employees’ tenure (years spent on current position) and perceived control over performing well. That means that perception of the control to perform certain task doesn’t relate to the experience of employee. However, we can clearly see that there is the high correlation between the perceived control over performing well and gender of employee (b=0.267, p<0.01).
To understand the differences in gender perception of perceived control and quality of relationship due to higher correlation coefficient between the gender and perceived control over performing well I run the Independent T-test, the results are reported below in Table 5.4:
Table 5.4 Independent T-test results Variables t N Mean N Mean PC 53 4.6415 51 5.14 -2.191* Employees total LMX 53 5.5795 51 5.58 -0.002 Management total LMX 68 5.8824 31 5.98 0.-742 Male Female Notes: n=106; **p<0.01; *p<0.05;
Overall we can observe that the perception of quality of relationships from the perspectives of employees and managers are almost equal for men and women. The different situation is for perception of perceived control over performing well, we can see that women have higher level of believe regarding their ability to perform the work task then men do. The difference in mean is 0.50 and p<0.05. It seems as quite unusual fact, however, our analysis shows that women in subordinated position are more confident in their job environment than men are.
To check for the support of first hypothesis whether there is the positive relationship between the quality of relationship between employee and his direct supervisor I run the hierarchical multiple regression. Even though all assumptions of the normality of distribution have been met, I have observed high level of negative skewness for perceived control over performing well (z-value -3.53), which indicates that higher proportion of the scores were at the higher end of the perceived control scale. To adjust for the violation of the normality distribution assumption I performed the bootstrap with bias corrected accelerated (BCa) with 1000 bootstrap iterations.
To identify whether such factors as gender, age, education and tenure of the employee could alter the beliefs regarding their ability to perform the work related tasks I executed the first step of the regression only for the control variables. Result of regression analysis used to test hypothesis presented below (Table 5.5).
Table 5.5 Regression results testing hypothesis 1
1 step 2 step 3 step
B B B 0.577* 0.579* 0.483 (0.085; 1.032) (0.086; 1.064) (-0.001; 0.952) 0.093 0.083 0.062 (-0.053; 0.258) (-0.07; 0.2550) (-0.088; 0.223) -0.002 -0.003 -0.002 (-.036; 0.033) (0.802; -0.034) (-0.035; 0.028) -0.002 0.003 0.008 (-0.048; 0.046) (0.919; -0.034) (-0.037; 0.059) 0.178 0.16 (0.212; -0.089) (-0.107; 0.059) 0.356* (0.051; 0.636) ∆R² 0.079 0.018 0.047 R² 0.079 0.097 0.144 F 1.907 1.892 2.443* Variable Employee gender Percieved control Employee education Employee age Employee tenure Employees total LMX Management total LMX
Notes: n= 103; **p<0.01; *p<0.05; B – bootstrap coefficient with the significance interval
The total fit of the model is 14,4% (R² = 0,144), that means model explain 14.4% of variance in depending variables. From the first step of running the regression we can notice that the variance in gender explain a significant amount of the variance in perceived control (7,9%). Indeed, the beta of the gender is high and significant (β =0.213, p<0.01). In the second step we can noticed that other additional contribution (6,5%) to the change in variance of perceived control is caused by the quality of the dyadic relationship (β=0.227, p<0.05). From the perspective of employee there is no any significant influence on variance of dependent variable. That means that the perception of employee of their ability doesn’t depend of their perception of the quality of the relationship with their manager. As the both variables ETLMX and perceived control to perform well are self -reported constructs the possible explanation to that could be that the employees prefer to recognize their own merit of their ability to perform well and not give the credit for somebody else (Kunz & Plaff, 2002).
The bootstrap confidence interval tells us that population value of beta for MTLMX (0.227) is likely to fall between 0.076 and 0.623, and because this interval doesn’t include zero we would
conclude that there is a genuine positive relationship between quality of leader-member exchange and perceived control to perform well. Also the significance associated with this confidence interval is p<0.05, which is highly significant. Similar relationship we can observe between the gender and perceived control (beta=0.213, p<0.05). Following this, I can assume that the firs hypothesis was partially supported.
To test for the moderation effect of the use of the performance indicators for incentive purpose I computed two new variables consisting of the product of the PIs and MTLMX, and PIs and ETLMX. To control for the effect of the moderator on the dependent variable, I also added a moderator variable itself as a predictor. To make sure that the differences come only as a result of the indicated independent variables, I run the first step of the regression for the controlling variables such as gender, age, education and tenure of employees. To facilitate the interpretation of the outcome of the regression I have centred the independent and moderation variables. The results of the regression are presented in Table 5.6.
Table 5.6 Regression results testing hypothesis 2
Step1 Step2 Step3 Step4
B B B B 0.577* 0.483 0.465 0.45 (0.062; 1.045) (-0.003; 0.941) (-0.049; 0.974) (-0.097; 1.015) 0.093 0.062 0.031 0.03 (-0.072; 0.26) (-0.103; 0.228) (-0.141; 0.194) (-0.149; 0.205) -0.002 -0.002 0.003 0 (-0.036; 0.032) (-0.035; 0.026) (-0.032; 0.028) (-0.035; 0.024) -0.002 0.008 -0.001 0.001 (-0.048; 0.054) (-0.039; 0.068) (-0.052; 0.065) (-0.052; 0.071) 0.356* -0.292 -0.285 (0.035; 0.7) (-0.068; 0.578) (-0.083; 0.599) 0.16 0.159 0.158 (-0.123; 0.44) (-0.111; 0.45) (-0.118; 0.448) 0.185 0.196 (-0.087; 0.444) (-0.103; 0.524) 0.101 (-0.34; 0.439) -0.003 (-0.263; 0.179) ∆R² 0.079 0.065 0.031 0.005 R² 0.079 0.144 0.175 0.18 F 1.907 2.443* 2.614* 2.049* Employee tenure Variables Percieved control Employee gender Employee education Employee age MTLMX_centered ETLMX_centered PIs_centered PIs_MTLMX PIs_ETLMX
The results of the table show that there is no moderation effect of use of the performance indicators for incentive purpose. That means that the influence of the quality of the leader-member exchange on perceived control of employee doesn’t change in the presence of performance indicators for incentive purpose. Even though Pearson correlation shows that there are significant association between PIs and perceived control to perform well, and between PIs and MTLMX, the relationship between quality of the leader-member exchange and perceived control of employees to perform well doesn’t change in function of the use of performance indicators. Therefore the second hypothesis was not supported.
Additional analysis
The result of main analysis shows no interaction effect of the use of PIs for incentive purposes on the relationship between quality of LMX and perceived control.
We also can notice that even though the gender was introduced only as a control variable and not as a variable of theoretical interest the level of perceived control is highly associated with the sex of the employees. Our analysis shows that subordinators women are more confident about their own ability to perform their tasks than men are.
Within the behaviour studies, it is often said that men have greater perception of control than women over their emotions and behaviour, suggesting that men are naturally stronger than women. (Why women rule the roost, S. A. Brenner, p.21), however in other studies (e.g., Instone, Major, &Bunker, 1983) women were found more confident and more competitive in their nature, and there is also study (e.g., Shockley-Zalabak, 1981) that show no differences due to the gender.
Therefore, it would be appropriate to check whether the gender variable could be the moderator in the relationship between LMX and perceived control. In order to run the linear regression with moderator gender, I create the dummy variable, Gender_women, in which men were chosen as reference category. The regression was conducted in 4 steps. In first step I have included the control variables with relation to employee data (age, tenure in current position, education); separate step was performed with introduction of only moderator variable (gender) as a predictor to control for direct effect on the dependent variable. The last step was run for
the interaction variable with new computed variables consisting of product of MTLMX and gender_women, and ETLMX and gender_women. Like in the mane analysis to avoid the misleading results of regression coefficient I perform the bootstrap with 1000 bootstrap iterations. The confidence intervals are reported in the parenthesis.
The results of the regression are shown in Table 5.7 below.
Table 5.7 Regression analysis testing interaction effect of gender e Predictors: (Constant), Employee tenure, Employee education, Employee age, MTLMX_centered, ETLMX_centered, Employee_women, Gender_MTLMX, Gender_ETLMX
Step1 Step2 Step3 Step4
B B B B 0.144 0.118 0.108 0.117 (-0.056; 0.361) (-0.083; 0.335) (-0.082; 0.32) (-0.077; 0.322) -0.02 -0.022 -0.018 -0.02 (-0.063; 0.02) (-0.065; 0.016) (-0.058; 0.016) (-0.06; 0.013) 0.023 0.033 0.022 0.027 (-0.027; 0.082) (-0.018; 0.101) (-0.029; 0.088) (-0.024; 0.089) 0.328 0.263 0.221 (-0.134; 0.646) (-0.212; 0.607) (-0.363; 0.625) 0.147 0.149 0.334 (-0.142; -0.418) (-0.127; 0.406) (-0.033; 0.702) 0.562* 0.542 (0.087; 1.002) (0.081; 1.008) 0.271 (-0.379; 1.066) -0.367 (-0.984; 0.24) ∆R² 0.043 0.051 0.053 0.019 R² 0.043 0.094 0.147 0.166 F 1.376 1.877 2.555* 2.164* Percieved control Employee education Employee age Employee tenure MTLMX_centered ETLMX_centered Employee_women Gender_MTLMX Gender_ETLMX Variables
Notes: n= 103; **p<0.01; *p<0.05; B – bootstrap coefficient with the significance interval
The results of the regression show no significant influence of the control variables. We can notice that the beta coefficient of gender as a predictor of the model is significant.
The interaction effect of gender variable was not found in our analysis. The level of influence of quality of exchange relationship on perception of the control doesn’t change in function of
perceived control over performing well (b=0.562, p<0.05). It means that according to our database the women score 0.5 points more on perceived control over performing well then men do. It proves again that women are found to be more confident in their job position then men are, even though the interaction effect of gender was not supported
6. Conclusion and limitation of the study
6.1 Discussion
This study is predicated on the assumptions that Leader-member exchange positively influences the perception of employee over performing well (perceived control), and that this relationship is moderated by the level of use the performance indicators for incentive purpose. Thus, the main objects of my research were:
1) discover how the quality relationship between managers and their employees can influence employee’s perception regarding his ability to perform relative tasks, 2) and to control whether there is the moderation effect of use of PIs for incentive purpose
on aforementioned relationship.
According to the theory of planned behaviour increase in perception of the employee of his ability to perform well leads to an increase in likelihood that the employee will act according to his intention. From the previous chapters we also know that perceived behavioural control increases with availability of resources and possibilities to have access to that resources. In the organizational environment the resources may be the physical materials but also knowledge, information, as well as human support. If the employee has easy access to these resources his perception of own control will eventually increase. On the other hand, according to the theory of LMX the high quality exchange relationship may provide the employee with open sharing of information, support from his supervisor, and trust in his ability (Graen and Cashman, 1975). These were the main factors for prediction of my first hypothesis. For the second hypothesis my intention was to discover whether the use of PIs for incentive purposes may influence the relationship between LMX and perceived control of the employee. As a reference I used the theory of the qualities of PIs for incentive purposes. According to this theory the behaviour of the employee tends to change in the moment he knows to be evaluated and rewarded in accordance with this evaluation.
Several important difficulties were encountered during the analysis as to operationalization of the LMX construct. Following the previous research in the field of LMX to avoid the common method bias of self-reported data I used two separate variables that explain the quality of the LMX from the perspective of employee and manager. It is important to notice that the
questionnaire content was the same for both parties. This may be the possible explanation for differences in the results of the analysis from the manager’s and employee’s perspectives. We obtained the significant relationship between the quality of LMX and perceived control over performing well from the manager’s perspective. The results of study are consistent with the past research on LMX. Many behavioural studies have shown that the high quality dyadic relationships positively influence the perception of individuals of their ability to perform and increase their level of self-efficacy (Wayne, Shore, and Liden, 1997; Graen, Novak, and Sommerkamp, 1982). However, no association was found from the perspective of the employee. Returning to the previous topic we know that employee and manager evaluate the quality of the relationship from different angles and take into consideration different aspects of the relationship. The employees tend to appreciate more the relationship in which they perceive more respect, affect and sensitivity, while managers consider the high-quality relationship in which they perceive high employee contribution, competence and loyalty (Dockery and Steiner, 1990 and Day and Crain, 1992).
Following this the possible recommendation for future research may be to use the different questionnaires for the supervisor and employee, trying to capture better the differences in perception of the quality of the exchange relationship in these two categories of individuals. Thus we can conclude that the first hypothesis was partially supported by this research. The analysis also indicates that gender has significant correlation with the perceived control of the employee. The results of the independent T-test show that women-subordinates are more confident in their job than men are. It somehow contradicts the common opinion according to which the men have more control and a higher perceived level of their ability. However, the previous studies demonstrate different results in behavioural research according to the difference in gender perception of self-efficacy (Shockley-Zalabak, 1981, Instone, Major, &Bunker, 1983). With my research I provide some insight into the influence of gender on the level of perceived control over performing well, with the recommendation to deepen this knowledge in more specific areas of organizational study.
The analysis of the interaction effect of use of PIs for incentive purposes on the LMX-perceived control relationship shows no support for the second hypothesis. This implies that intensity of use of performance indicators does not change the level of influence of the exchange relationship on the perceived control of the employee. Contrary to expectation, the use of PIs has no interaction effect. Nevertheless, these findings provide the insight into managerial
studies which suggest to exclude the possibility of use of the PIs as instruments to increase the perception of employees regarding their ability to work well. Because this fact has not been clearly addressed in the literature, this possibility suggests another avenue for future research. Having a significant relation of gender to the perceived control the additional analysis has been computed with the purpose to determine whether there is a moderation effect of the gender on the LMX-perceived control relationship. The result shows that gender does not moderate the aforementioned relationship. However, we have noticed a significant direct correlation of gender to the perceived control.
Research exists on how LMX influences the performance (Wayne, Shore, and Liden, 1997) and productivity (Graen, Novak, and Sommerkamp, 1982), however this study goes beyond the mere measurement of performance. The intention of this research was to deepen understanding of the determinants and stimuli that drive employees to perform and be pro-active and persistent in their actions. I attempted to understand the changes in the behavioural constructs in various working environments that are differentiated by the quality of exchange relationships.
Since LMX is positively correlated with the performance (B. Groen, 2015) it is important for the organization to initiate sound leadership training to emphasize the importance of human relationship skills, reciprocal exchange of information between employees and their managers, cooperation for development of goals in order to enhance interpersonal communications and exchange relationships. In practice the result of this study highlights the importance of high-quality relationships between employees and supervisors and encourage the development of group interactions as a practical area for building prospective relationships. The team-building event, development seminar and thematic workshops would be very helpful for development of an effective working environment which will contribute to the increased level of trust, personal connection, understanding of needs and problems that need to be addressed and finding common effective solutions (T. A. Scandura, G.B. Graen. 1984). Since the managers play the key role in adopting and maintaining an organizational culture of trust (St-Pierre and Holmes 2010) it would be appropriate for managers to act as role models and initiate a positive influence on their subordinates.
6.2 Limitations
Limitations of the study can arise from the characteristics of the sample. It may not be representative for other populations. Results can be different depending on the type of person, the type of sector in which the person operates and organizational culture that defines the patterns of the relationship between manager and employees.
Additionally, there may arise concerns related to the quality of questionnaire data, which may result in mono-method bias (Kuvaas 2009; Sharma et al, 2004). Although it seems obvious at first glance, this study clearly shows the necessity for improving theorization about LMX and measurement techniques. As was anticipated in the literature part, that previous studies show some basic definitional and validity problems of the LMX construct (C.A. Schriensheim, 1999, S. l. Castro, 1999). Scholars continue to define the LMX as a quality of the relationship between employee and supervisor, while at the same time approaching different subdimensions and subcontent of the construct.
One limitation of the study was the use of the same questionnaire for both manager and employees to assess the quality of the LMX relationships. The perception of the quality of the dyadic relationship could be based on different factors in relation to the specific needs of the participants. It also may be the case (with managers’ survey data) of attribution bias when one attributes the positive events to oneself and negative events to the external factors.
Although the survey design was made in the way to reduce potential common method bias (Podsakoff et al., 2003) there could be the possibility that the bias still exists taking into account the behavioural nature of the research. According to Podsakoff some common-methods bias may result from the fact that the respondent providing the measure of the predictor and criterion variable is the same person. In my case the employees provided measures to both variables, namely LMX and perceived control over performing well. Therefore the results of my analysis could be affected by common method bias. Also the sensitivity of the questions and transient mood state of the respondents may produce artefactual covariance in self-reported measures. Significant limitation may be the fact that in our sample all organizations were using the PIs and there was no possibility to control whether the organization without the use of PIs for incentive purposes would produce any different effect on the LMX and perceived control relationship.
6.3 Conclusion
We can see how the different perception of LMX (managerial and employee) have different correlation with the perceived control to perform well of the employee. With regard to these findings and limitations of the current study there are possible recommendations for future research. Longitudinal study might provide insight into how dyadic relationships grow over time. Since the level of performance of the employee depends on many behavioural factors (B. Groen, 2015) it would be appropriate for practitioners and academics to expand the theoretical background of behavioural antecedents that lead directly to organizational performance and examine potential moderator and intervening factors in the relationship between LMX and behavioural beliefs of employees. One of the possible directions for further research could be to study LMX influence on different behavioural determinants of performance, for example, taking into the span of control or dyadic tenure. Organizing these findings may help direct future research efforts.
Hopefully my study will serve to further stimulate the interest in the field of leader-membership exchange relationships and behavioural control to improve the quality of future knowledge and to explore this domain deeper.
