MSc in Business Studies - Leading and Managing People
Master Thesis - Linsy Freeke – 6137679 – March 2014
How HRM relates to employee turnover intentions
A mediated moderation model
Supervisors
Dr. Corine T. Boon
Dr. Deanne N. den Hartog
Department of Economics and Business
University of Amsterdam
Abstract
This quantitative multi-source study examines the relationship between HRM and employee turnover intentions. First, based on the ability, motivation and opportunity (AMO) theory and social exchange theory it is proposed that HRM is negatively related to turnover intentions. Second, based on the attraction, selection, attrition (ASA) theory, it is proposed that Person-Organization fit (PO-fit) mediates the relationship between HRM and turnover intentions. Third, also based on social exchange theory it is proposed that trust in the manager moderates the relationship between HRM and PO-fit, which in turn affects turnover intentions. This study is conducted through a survey in a broad set of companies in different sectors and countries. Results from 371 usable manager-employee dyads supported the mediated
moderation model. Person-Organization fit fully mediated the relationship between HRM and turnover intentions. Furthermore, a significant indirect effect is found from HRM to turnover intentions via PO-fit for medium and high trust in the manager, but not for low trust in the manager. Implications and directions for future research are discussed.
Keywords: HRM, PO-fit, Turnover intentions, Trust in the manager, Moderation, Mediation
Table of contents
1. Introduction ... 4Error! Bookmark not defined. 2. Literature review ... 7Error! Bookmark not defined. 2.1 HRM ... Error! Bookmark not defined. 2.2 Employee turnover intentions ... Error! Bookmark not defined. 2.3 HRM and employee turnover intentions ... Error! Bookmark not defined. 2.4 The mediating role of Person-Organization fit ... Error! Bookmark not defined. 2.5 The moderating role of trust in the manager ... Error! Bookmark not defined. 3. Method ... Error! Bookmark not defined. 3.1 Procedure and sample description ... Error! Bookmark not defined. 3.2 Description of measures ... Error! Bookmark not defined. 3.3 Analytical strategy ... Error! Bookmark not defined. 4. Results ... Error! Bookmark not defined. 4.1 Data preparation ... Error! Bookmark not defined. 4.2 Analyses ... Error! Bookmark not defined. 5. Discussion ... Error! Bookmark not defined. 5.1 Conclusion ... Error! Bookmark not defined. 5.2 Theoretical and managerial implications ... Error! Bookmark not defined. 5.3 Practical implications ... Error! Bookmark not defined. 5.4 Limitations and directions for future research ... Error! Bookmark not defined. 6. References ... Error! Bookmark not defined. 7. Appendix ... Error! Bookmark not defined. Appendix 1: Introduction letter for respondents ... Error! Bookmark not defined. Appendix 2: Survey Manager... 47Error! Bookmark not defined. Appendix 3: Survey Employee ... Error! Bookmark not defined. Appendix 4: Output MODMED Syntax ... Error! Bookmark not defined.
List of figures
Figure 1. Research Model.
Figure 2. Withdrawal decision process (Mobley, 1997). Figure 3. Mediated moderation model (Preacher et al., 2007).
Figure 4. The moderating role of trust in the manager between HRM and PO-fit.
List of tables
Table 1. Normality test.
Table 2. Descriptives, correlations, and reliabilities of the study variables. Table 3. Multicollinearity analysis.
Table 4. Multiple regressions: HRM and turnover intentions mediated by PO-fit. Table 5. Results of hierarchical regression analysis.
1. Introduction
When employees leave the organization, the impact can be substantial. Turnover is directly linked to an increase in employee recruitment and training costs and low levels of job satisfaction (Cho, Johanson & Guchait, 2009). Furthermore, turnover intentions is the strongest predictor of actual turnover (Tett & Meyer, 1993; Carmeli & Weisberg, 2006). In order to prevent actual turnover, it is of great importance for an organization to remain their human capital and to influence the turnover intentions of their employees. But what factors are influencing turnover intentions of employees? And how can an organization retain their employees?
In recent years, in the field of strategic human resource management (SHRM), scholars and practitioners have increasingly recognized that human assets are the most
important assets for an organization (Cascio, 2006). The knowledge, skills, abilities and other characteristics (KSAOs) of the employees that are embedded in the firm’s culture and history have the potential to be valuable, rare, inimitable and non-substitutable due to social
complexity and causal ambiguity (Barney & Wright, 1998). An organization should exploit these sources of sustainable competitive advantage in an optimal way to increase employee performance (Grant, 1996 in Wright, Dunford & Snell, 2001). Grant (1996) stressed the importance of effective human resource management (HRM) for organizational and employee performance. Besides, Batt and Valcour (2003) emphasize the importance of employee
turnover intentions, because employees and the knowledge of employees are irreplaceable assets and the costs of turnover are high. Therefore, it is not only important to exploit human resources in an optimal way, but also to retain human assets (Noe, Hollenbeck, Gerhart & Wright, 2012). There is a large role for human resource management to attract, develop and retain human capital (Combs, Liu, Hall & Ketchen, 2006).
In recent decades, research has focused on the relationship between HRM and turnover intentions (Boselie & Van der Wiele, 2002; Luna-Arocas & Camps, 2007; Mostafa & Gould-Williams, 2014). HPWPs have a clear negative relationship with organizational turnover intention rates (Huselid, 1995). However, there is a research gap in the understanding of the mechanisms linking HR practices to employee turnover intentions. This results in calls in the SHRM literature for research attention in this area (Mostafa & Gould-Williams, 2014). The contribution of this study to the SHRM literature is twofold. First, the current study aims to explore further the so called ‘black box’ between HRM and employee performance, in specific employee turnover intentions. In other words, the current study will address the
research gap by investigating how and when (under what circumstances) the relationship between HRM and employee turnover intentions occurs. The aim is to extend current research by examining PO-fit as mediator in the relationship between HRM and employee turnover intentions (Figure 1 depicts the research model, p.6). According to Boon, Den Hartog, Boselie and Paauwe (2011), human resource practices as a coherent system can communicate
organizational values to employees. This increases congruence between values of employees and their organizations and in turn establish PO-fit. Through the process of attraction,
selection and attrition a certain kind of people are attracted and selected, of which it is most likely that their values are in congruence with the values of the organization. Therefore, it is proposed that employees are less likely to leave the company when there is a
person-organization fit. Furthermore, to explore the black box further trust in the manager is expected to function as a moderator. Based on social exchange theory, it is expected that trust in the manager will have a positive influence on the indirect relationship between HRM and turnover intentions through PO-fit. Employees with high trust in the manager experience strong individual consideration from their manager. Therefore, employees tend to be strongly affiliated with and loyal to their organization. Enhanced trust in the manager will decrease employees intentions to leave the organization, because employees are willing to maintain membership with the organization (Mayer & Davis, 1999).
Second, this study adds value to the existing literature, because the data is collected by a multi-source survey reported by employees as well as managers. HRM can influence the turnover intentions of employees, but HRM practices are not always received as intended. Managers are responsible to communicate HR practices as how they are actually
implemented. However, employees can receive and interpret the HR practices in different ways (Nishii & Wright, 2008). For this reason, HRM is rated by managers and in this way measured as how they are implemented. These HR practices rated and communicated by managers affect the attitudes and behaviors of employees. To get a clear picture of how these HR practices are received by employees, PO-fit, trust and turnover intentions will be
measured by the employees themselves.
This study will first discuss the primary theoretical mechanisms predicting a negative relationship between HRM and turnover intentions. It will then describe the mechanism PO-fit, through which HRM indirectly affect turnover intentions. Furthermore, trust in the manager will be discussed as a potential moderator. Following a description of the research methodology is presented, three hypotheses are tested using regression analyses and the
MODMED syntax of Preacher, Rucker and Hayes (2007). The final section of this study provides an analysis of the results and discusses the implications for science, managerial practice and future research.
2. Literature review
2.1 HRM
Human resource management as the independent variable of this study refers to ‘the policies, practices and systems that influence employees’ behavior, attitudes and performance. HRM practices include analysing and designing work, determining human resource needs (HR planning), attracting potential employees (recruitment), choosing employees (selection), teaching employees how to perform their jobs and preparing them for the future (training and development), rewarding employees (compensation), evaluating their performance
(performance management), and creating a positive work environment (employee relations)’ (Noe et al., 2012, p. 5).
It is important that high performance work practices (HPWPs; in this study also referred to as HRM) are constructed as integrated and coherent systems of mutually
reinforcing practices. These powerful combinations of HR practices are designed to enhance the skills and efforts of employees in organizations. Such coherent systems of HR practices are likely to better support performance outcomes than individual HR practices (Boselie, Dietz & Boon, 2005; Combs et al., 2006). According to the ability, motivation and
opportunity (AMO) model employees will perform well when they are able to do so. If they possess the necessary KSAs, they have the ability to perform the job. Besides, employees will perform well when they have the motivation to do so. They will do the job, because they want to and are adequately paid. Finally, employees will perform well when their work
environment provides the necessary support, resources and avenues for expression and they have the opportunity to participate (Appelbaum, Bailey, Berg & Kalleberg, 2000). HPWPs enhance the three main components of the AMO model to perform behaviors consistent with organizational goals and strategy. Specifically, HPWPs increase employees’ KSAs, empower and motivate employees to use their KSAs for organizational benefit (Combs et al., 2006). HPWPs enhance individual, unit and organizational performance by different processes. For example, recruiting and selection procedures are important to get the right employees. Subsequently, feedback, training, job design, and performance appraisal linked with compensation will enhance their KSAs and motivation. Furthermore, practices such as pay for performance and internal promotion opportunities can also increase the motivation of employees. Commitment to the organization can be increased by employment security, flexible work hours and high overall compensation. Finally, career management practices
provide employees with opportunities to develop themselves (Combs et al., 2006).
This study includes a combination of human resource practices, which reflect one high performance human resource system. The HRM scale used in this study contains ability enhancing, motivation-enhancing and opportunity-enhancing practices. The HPWPs that enhance abilities are formal selection tests, hiring selectivity, training opportunities and structured interviews. HPWPs that enhance motivation include: formal performance evaluation mechanisms, high pay, merit-based promotion systems and rewards based on individual and group performance. Finally, opportunity-enhancing practices include: formal participation processes, regular communication and information-sharing efforts and autonomy in work-related decision making (Boselie, 2010; Kehoe & Wright, 2013).
Research about HPWPs has largely focused on the impact on organizational
performance, while there is less attention for the impacts on the individual employee and their attitudes and behavior (Mostafa & Gould-Williams, 2014). Findings of previous studies suggest that HRM is positively related to employee outcomes like job satisfaction,
organizational commitment, citizenship behaviors and negatively related to intention to quit (Boon et al., 2011; Gould-Williams & Gatenby, 2010). Recently, research paid more attention to employee outcomes in order to explain the relationship with organizational performance (Boselie, 2010; Kehoe & Wright, 2013). According to Campbell (1999), it is critical to explain the relationship between HR practices at the organizational level and withdrawal at the individual level. He further argues that there is limited evidence from empirical studies linking HR practices to individual turnover intentions. It is important to investigate turnover intentions to prevent recruitment costs and to keep levels of job satisfaction high. As
discussed in the introduction, turnover is directly linked to an increase in employee recruitment and training costs and low levels of job satisfaction (Cho et al., 2009).
Furthermore, turnover intentions is the strongest predictor of actual turnover (Tett & Meyer, 1993; Carmeli & Weisberg, 2006). In order to prevent actual turnover, it is of great importance for an organization to remain their human capital and to influence the turnover intentions of their employees. For this reason turnover intentions is chosen as specific employee outcome of HRM. But how can an organization retain their employees? And what factors are influencing turnover intentions of employees? These questions will be answered in the next sections. The first section will explain turnover intentions, followed by a section in which the relationship between HRM and turnover intentions will be discussed in more depth.
2.2 Employee turnover intentions
This section is about turnover intentions as the dependent variable in this study. Turnover intentions refer to “the conscious and deliberate willingness and intention of an employee to leave the current organization” (Tett & Meyer, 1993, p. 262). In other words, turnover intentions can be described as the subjective estimation of an employee regarding the
probability of leaving the organization in the near future. The words ‘turnover intentions’ and the concept ‘intention to leave’ can be used interchangeable.
There is a comprehensive explanation for the psychological process underlying employee withdrawal and turnover. According to the withdrawal decision process of Mobley (1977), one of the consequences of dissatisfaction are thoughts about quitting and leaving the organization. These thoughts, in turn, stimulate considering to search for new jobs and the costs of quitting. This will be followed by the behavioral intention to search for an alternative, in turn an actual search, and if there are alternatives, an evaluation of these alternatives is initiated. The employee will compare the current job and the alternatives and if the comparison favours one of the alternatives, the employee will be inspired by a behavioral intention to quit. Finally, this will be followed by actual withdrawal (see the different steps of the withdrawal decision process in Figure 2, p.10).
Hom and Griffeth (1991) built further on the extensive withdrawal decision process of Mobley (1977), and argue that turnover intentions can be operationally split up in three
elements of a cognition process. These three elements are thoughts of quitting, the intention to search for another job elsewhere and the intention to quit. Intentions to quit is considered as the last part in the withdrawal cognition process. Carmeli and Weisberg (2006) show that intent to leave is the strongest predictor of actual turnover. For this reason it is of great importance that organizations know how to influence turnover intentions, so they can prevent actual turnover.
Figure 2. Withdrawal decision process (Mobley, 1997).
2.3 HRM and employee turnover intentions
HRM affects the attitudes and behaviors of employees. Specifically, turnover intentions of employees is one of the behaviors that can be influenced by HRM (Huselid, 1995). HRM can be linked to the withdrawal decision process. Besides positive employee outcomes of HRM, there are possible negative effects of HRM on employees, which can cause feelings of dissatisfaction. For example, employees are not satisfied with the opportunities for training and development and are not getting feedback on a regular basis, which they did expect. Besides, employees can experience high levels of stress, because of work intensification. In this way, they are experiencing that there work-life balance is out of balance. These feelings of dissatisfaction lead to thoughts about quitting and can increase turnover intentions (Boselie et al., 2005).
This shows that it is essential to design a human resource system that positively influence the attitudes and behaviors of employees, so that turnover intentions will decrease. The relationship between HRM and turnover intentions can be explained by different theories. This will be done in the following paragraphs by referring back to the ability, motivation and opportunity (AMO) theory and explaining the social exchange theory. To start with the AMO theory, feedback through HPWPs foster learning and increase self-efficacy. This will enhance the feelings of an employee that they are convinced about their own abilities, and that they believe they are able to do the job. Autonomy in work-related decision-making through HPWPs can enhance employees’ attitudes of task significance. If employees perceive their tasks as significant, they are more motivated to perform the job. Opportunities for employee participation in job design, gives employees the feeling that their voice counts in the
organization and that they can make a difference with their work. Organizations that provide employees with these resources and support are more motivated to make meaningful
contributions in the employment context. The AMO model underpins the relationship
between HRM and turnover intentions, because HRM can enhance the three main components of the AMO model (ability, motivation and opportunity). This ensures that employees feel a strong bond to the organization and are motivated to stay at the organization. Also, meta-analysis of research in organizational behavior demonstrate that turnover is lower among employees who perceive human resource incentives of high relative pay, employment security, high pay satisfaction and career-development opportunities (Combs et al., 2006). Employees who experience greater autonomy and control over managing their work-family interface will show less likelihood of quitting their current employment. So, work designed to increase decision-making autonomy will lower turnover intentions.
Besides the AMO theory, the relationship between HRM and turnover intentions can be explained on the basis of the social exchange theory. High performance HR systems rely on the creation of a mutual investment-based relationship between the manager and employee. In this relationship a manager invests in workforce skills and opportunities for the employee. In turn, a manager expects employees to be qualified and motivated to make valuable work related investments in the organization. Therefore, the employee-manager relationship can be seen as a social exchange. Social exchange is defined as cooperation between two or more parties for mutual benefit. According to Blau (1964), social exchanges involve unspecified obligations in which individuals do others a favour, for which they expect some future return. The form of return and the time of return are often unclear. Social exchanges may also
involve reciprocity, which is defined as the norm that obligates the recipient of the benefit to repay the donor in some way. A strong affective bond towards the organization is a likely attitudinal consequence of a positive mutual social exchange relationship between the manager and subordinate. In this study, the interest is only in the behavior that reflects this bond. Intent to remain with the organization represents the behavioral consequence of employees’ affective bond to the organization. This can be reflected in a desire to see the organization succeed in its goals and a feeling of pride at being part of the organization (Cohen, 2003).
Specifically, social exchange theory suggests that individuals are likely to participate and invest in rewarding relationships (Blau, 1964). High-performance HR practices, if implemented effectively, are likely to cause employees to perceive that their exchange relationship with the manager is characterized by a supportive environment. This is based on investments in employee skills, regular unbiased performance feedback, availability of fair and attractive rewards for performance and mutual efforts towards meaningful goals (Wright, Gardner, & Moynihan, 2003). In return, employees are likely to feel an obligation to the goals of the manager and the organization. So, they develop an affective bond with the manager and organization itself based on a sort of obligatory reciprocation in their exchange relationship (Cohen, 2003; Sun, Aryee and Law, 2007). Employees who are more involved in the organization are likely to more naturally behave in ways that reflect this affective bond. In particular, these employees are likely to act in ways that are in the best interest of their organization. Furthermore, strongly involved employees managed with a high-performance HR system are likely to feel both eager and obliged to contribute to organizational goals (Blau, 1964; Cohen, 2003). They are probably more willing to stay at the organization than less involved employees. This is because of a few reasons. First, employees want to contribute to organizational effectiveness only as they continue their employment relationship. In other words, employees’ desire to help achieve organizational goals aligns with an intent to remain with the organization. Second, employees will leave an organization if they believe their incentives match no longer the contributions they make. The desire to leave represents a form of withdrawal. This can be due to a lack of emotional attachment to the organization and its goals. The central assumption of social exchange theory is that HRM practices are viewed from an employee perspective as a personalized commitment to them by the organization. This is reciprocated back to the organization by employees through their intention to stay working at the organization (Mayer & Davis, 1999).
To conclude, HPWPs are a synergistic whole of performance and commitment enhancing human resource practices, wherein each practice can contribute to a higher organizational and employee performance (Combs et al., 2006). Turnover intentions can be explained by the use of the withdrawal decision process and is the intentions of an employee to leave the current organization. The relationship between HRM and turnover intentions can be explained by AMO theory and social exchange theory. HPWPs that stimulate employee ability, employee motivation and employee opportunity to participate (AMO) seem to create a strong bond between the employee and organization. In turn it has a negative effect on
employee turnover intentions. In order words, HRM is negatively related to employee turnover intentions, because employees are less likely to leave the organization with an effective HR system. According to the social exchange theory, a strong affective bond towards the organization is a likely consequence of a positive mutual exchange relationship, which will result in lower turnover intentions and willingness to stay at the organization.
Hypothesis 1: HRM is negatively related to employee turnover intentions.
2.4 The mediating role of Person-Organization Fit
In the previous section a negative relationship between HRM and turnover intentions is proposed. However, the mechanisms through which HRM affect employee outcomes as turnover intentions still need more research (Mostafa & Gould-Williams, 2014). If an
organization wants to hold their employees, it is important that they are aware through which mechanisms HRM is affecting the turnover intentions of their employees. The current study will address this research gap by analysing a mediator. Preacher et al. (2007, p. 186) define mediation as ‘an indirect effect which occurs when the causal effect of an independent variable (X) on a dependent variably (Y) is transmitted by a mediator (M)’. In this study PO-fit is proposed as a mediator between HRM and employee turnover intentions. It is expected that HRM can have positive effects through enhancing value congruence in establishing PO-fit (Boon et al., 2011). Furthermore, Cable and Judge (1996) show that person-organization PO-fit is useful to make better predictions of employee’s attitudes and behavior. Researchers propose that person-organization fit is a good predictor of turnover intentions (Cable & Judge 1996; Cable & DeRue, 2002; Kristof, 1996).
From an organizational behavior perspective, person-organization fit (PO-fit) refers to congruence between specific characteristics of an organization and the people in it. Kristof
(1996, p. 4) defines PO-fit as ‘the compatibility between people and organizations that occurs when: (a) at least one entity provides what the other needs or (b) they share similar
fundamental characteristics or (c) both’. Specifically, PO-fit perceptions refer in this study to judgments of congruence between employee’s personal values and the organization’s values. Schneider (1987) proposes that individuals may be attracted to organizations they perceive as having values similar to their own and which they are less likely to leave. Values provide the starting point, with socialization processes acting as complementary means to insure PO-fit. The attraction-selection-attrition (ASA) framework of Schneider (1987) helps explain the means by which HRM affect fit between employees and their organizations and in turn turnover intentions. The fundamental idea of the ASA framework is that organizations attract, select and retain people whose personal values fit the organization. People are attracted to different kinds of organizations based on their pre-entry beliefs of the organization’s principal values and goals. Thereafter, organizations choose, through formal and informal selection strategies, people who fit their values and goals. Finally, the theory proposes that where individuals do not fit the organizations core values, they will tend to leave (self-selection process). It may reflect unfulfilled expectations on the part of the employee or errors of judgment in the selection process of the organization. In sum, it is expected that employees leave an organization in which they perceive a mismatch of values (Schneider, 1987). If employees believe that their values are in congruence with the values of the organization and the values of the other employees working in that organization, they
probably feel involved with the broader mission of the organization. If person-organization fit is achieved, employees are likely to define themselves (at least partly) in terms of their
organization (Saks & Ashforth, 1997). When employees perceive their values in congruence with the values of their organization, they will be more likely to attribute positive motives to the organization’s behaviors and actions toward them. Finally, Saks and Ashforth expected that person–organization fit perceptions affect employee turnover intentions. This can be explained by the extent that individuals perceive the values of their organization reflect their own identity, which should result in a strong bond to the organization. This, in turn, makes leaving more difficult even if better tangible rewards are available elsewhere. Furthermore, if the values of the employee are in congruence with the values of their colleagues, employees find it easier to communicate and work with others within the organization. This strengthens their bond with their colleagues in the organization and in turn makes it more difficult for the employee to leave (Cable & DeRue, 2002).
There are many studies that have examined the effects of PO-fit on employee
outcomes, although fever have considered how PO-fit can be established (Mostafa & Gould-Williams, 2014). Boon et al. (2011) argue that HRM plays a role in establishing PO-fit, because HPWPs increase congruence between values of employees and their organizations. As a package HR practices communicate organizational values to employees. HPWPs are one of the major factors that help match employees with their organization. HPWPs as selection, reward systems, promotion and training and development consistently communicate
organizational values and expectations to employees, which in turn facilitate greater PO-fit (Boon et al., 2011). According to ASA theory, positive perceptions of PO-fit result in
employees feeling more satisfied with their jobs and the organization they are working for and therefore not willing to leave. Managers should endeavour to use HPWPs to facilitate greater congruence between employees and organizations in order to achieve improved employee attitudes and behaviors in the area of turnover intentions.
Mostafa and Gould-Williams (2014) show that high performance human resource practices (HPHRPs) are negatively related to employee turnover intentions and that this relationship occurs via the mechanism of person-organization fit. Results of the study of Mostafa and Gould-Williams (2014) show that HPWPs are positive related with PO-fit. Empirical results support that congruence between individual and organizational values are associated with positive affect and a high likelihood of their staying in their jobs (O’Reilly et al., 1991). O’Reilly et al. showed a significant negative relation between PO-fit and intentions to leave an organization (r = -.37, p<.01). This relationship indicates that respondents with high PO-fit as measured at time 1, when respondents originally entered their firms, is
associated with low intent to leave at time 2, a year later. Thus, PO-fit positively predicts the probability of a person’s staying with a firm. Failure to fit may reduce employees’ satisfaction and commitment and increase the likelihood of their leaving (O’Reilly et al., 1991).
Furthermore, fit between an individual’s preference for a particular culture and the culture of the organization the person joins is related to turnover intentions (O’Reilly et al., 1991). Finally, HPWPs not only lead to better fit between employees and organizations, but it is also associated with desired employee outcomes like employees who do not have the intention to leave the organization.
Even though, there is a high need for further research to confirm the findings of extant literature. In line with Schneider’s (1987) attraction–selection–attrition model, it is necessary to improve conceptual models of turnover by modelling person–organization fit as an
antecedent. Continuing, Mostafa and Gould-Williams (2014) stresses the importance for further research on how PO-fit can be established.
Hypothesis 2: PO-fit mediates the relationship between HRM and turnover intentions.
2.5 The moderating role of trust in the manager
The next aim is to expand our knowledge of when HRM is related to turnover intentions in a mediated moderation model. The sections above propose that the relationship between HRM and turnover intentions is mediated by PO-fit. Besides the intermediary mechanism PO-fit that explains the relationship between HRM and turnover intentions, a second question in this study is in which situations HRM is most strongly related to turnover intentions via PO-fit. To open the black box further, it is of critical interest to determine whether or not the mediation effect of PO-fit remains constant across different values of a moderator, because the
magnitude of an indirect effect may be dependent upon a moderator. In this study trust in the manager will function as moderation. So, it is proposed that the indirect effect may be dependent upon trust in the manager. The degree of trust in the manager will differ among employees. High trust in the manager results in a strong confidence in the abilities of the manager to manage effectively HRM. These employees experience congruence between their values and the values of their manager. This in turn will lead to a strong bond with the
organization and a strong motivation to stay at the organization. So, it is proposed that high trust in the manager will reinforce the relationship between HRM and PO-fit and in turn turnover intentions. On the other hand, employees with low trust in the manager are not convinced about the managerial abilities to manage HRM. Furthermore, these employees experiences a lack of concern for employee wellbeing of their manager and they have the feeling that there is no congruence between their values and the values of the manager. This have a negative effect on the bond and loyalty of the employee to the organization and the employee is less likely to stay at the organization. Low trust in the manager has a detrimental effect on the relationship between HRM and PO-fit and in turn turnover intentions. For this reason, trust in the manager as potential moderator will be further discussed in this section. Preacher et al. (2007, p. 191) refer to moderation when ‘the strength of the relationship between two variables is dependent on a third variable. The third variable, a moderator (W), interacts with X in predicting Y’. Applied on the current study, it means that trust in the manager as moderator interacts with HRM in predicting turnover intentions through PO-fit.
Why is trust in the manager important and how can trust between employee and manager arise? To answer this question, the relationship between employee and manager will be discussed in more detail.
The role of front line managers (FLMs) is mostly neglected in studies about the link between HRM and performance. However, these managers are the persons who are
responsible for the implementation of HR practices. Moreover, managers are in most situations the deliverers of HR practices. Besides, the quality of the relationship between employees and their immediate manager is liable to influence (positively or negatively) perceptions of HR practices and work climate. In turn, employee perceptions of the HR practices as well as their relationship with their FLM influence employee commitment to the organization. According to Purcell and Kinnie (2006), the responses of employees to HR practices are at the heart of all HRM-performance models, because it is the link between employee reactions and their subsequent behavior which is critical. Line managers need to be trusted by employees in order to have influence on the attitudes and behaviors of employees (Purcell & Hutchinson, 2007).This stresses the importance of the role of the manager and of the relationship between employees and their immediate manager. Why trust in the manager is important and what trust is will be further explained in the next paragraphs.
Trust within organizations is becoming of increased importance to management literature, due to a growing body of evidence linking trust to various performance outcomes, such as organizational citizenship behavior (OCB) (Colquitt, Scott, & LePine, 2007), job satisfaction (Dirks & Ferrin, 2002), knowledge sharing (Levin & Cross, 2004), organizational commitment (Liu & Wang, 2013), and decreased intentions to quit (Connell, Ferres &
Travaglione, 2003; Dirks & Ferrin, 2001; Davis, Schoorman, Mayer, & Tan, 2000). Empirical studies have demonstrated the importance of supervisor-employee relationships on job
satisfaction (Volmer, Niessen, Spurk, Linz & Abele, 2011; Harris, Wheeler, & Kacmar, 2009), organizational commitment (Eisenberger et al., 2010), and turnover intentions (Harris et al., 2009). From a strategic perspective, empirical evidence suggests that trust in the manager-subordinate relationship can even bring a significant competitive advantage to businesses, through reduced turnover (Barney & Wright, 1998; Davis et al., 2000). Several studies found a significant negative relationship between trust in the manager and intentions to leave (Sun et al., 2007; Mayer, Davis & Schoorman, 1995). These empirical results show the importance to include trust in the manager as moderator in this study. Specifically, this study will focus on trust in the manager from an employee perspective.
The most promising theoretical foundation for understanding the development of trust is the framework of Mayer, Davis and Schoorman (1995). They define trust as ‘the
willingness of a party to be vulnerable to the actions of another party based on the expectation that the other will perform a particular action important to the trustor,
irrespective of the ability to monitor or control that other party’ (Mayer et al., 1995, p. 712). In this case, the other party is the manager of a particular employee. The employee is the trustor and the manager can be seen as the trustee. According to Mayer et al. (1995), trust is comprised of three factors, namely ability, benevolence and integrity.
First, ability is ‘that group of skills, competencies and characteristics that allow a party to have influence within some domain’ (Mayer et al., 1995, p. 712). This means both the formal and informal influence of managers, as well as their perceived competences and skills. Employees can trust their manager if they feel confident about their manager’s skills and if their manager has much knowledge about the work that the employee needs to do.
Furthermore, employees can trust their manager if the manager is known to be successful at the things he/she is doing.
Second, benevolence is ‘the extent to which a trustee is believed to want to do good to the trustor, aside from an egocentric profit motive’ (Mayer et al., 1995, p. 712). Benevolence is explained as employees’ perceptions concerning the extent to which managers value employees’ contributions and care about their well-being. For example, employees perceive their manager having benevolence, if they maintain an accurate and fair appraisal system through which employees get developmental feedback. Actions of the manager that enhance an employee’s opportunities for success will be perceived as demonstrating benevolence. It is about the degree to which a manager is committed to its employees. An employee experiences their manager as having benevolence, if an employee believes a manager cares about the employee’s interest. Employees experience support from their manager if they have the feeling that their manager is listening to them, if the manager equally treat his/her
subordinates and if the manager give the necessary resources for the employee to conduct the job. For this to occur Whitener, Brodt, Korsgaad and Werner (1998) advise that managers must be encouraged to make the first move and that the initiation of this process, is the challenge and the responsibility of managers.
Third, integrity is defined as ‘the trustor’s perceptions that the trustee adheres to a set of principles that the trustor finds acceptable’ (Mayer et al., 1995, p. 712). This means not only that a manager show values that the employee sees as positive, but also that the manager
acts in a way that is consistent with the values. Moreover, employees trust their manager if the manager has a strong sense of justice. Besides, trust can be created if employees are
convinced that their manager will stick to his/her word and tries hard to be fair in dealing with others. In every situation the perceptions of a trustee’s ability, benevolence and integrity form the basis of trust. The combination of these perceptions have major influence on the level of trust for the trustee.
In organizational settings, the relationship between a manager and subordinate is of great importance to understand the attitudes and behaviors of employees (Jablin, 1979; Napier & Ferris, 1994). Researchers use social exchange theory as a theoretical foundation for
understanding employee-manager relationships. The relationship between HRM and turnover intentions is explained on the basis of social exchange theory. In addition, the link between employees’ experiences of their manager and their subsequent behavior can be explained by social exchange theory. A social exchange is based on undeclared obligations and trust. Social exchanges may also involve reciprocity, which is defined as the norm that obligates the
recipient of the benefit to repay the donor in some way. A high quality exchange relationship is related to feelings of trust and negatively related to intention to quit (Vecchio & Gobdel, 1984).
In this study it is proposed that trust in the manager will have a positive influence on the mediated relationship between HRM and turnover intentions by person-organization fit. The current study will test whether mediation exists for different conditional values of trust in the manager. This study propose that the relationship between HRM and PO-fit will be weaker for low trust in the manager than for high trust in the manager. Employees with high trust in the manager are likely to experience clear managerial communication of
organizational values (Mayer & Davis, 1999). Besides, these employees have the feeling that the values of their manager are in congruence with their own values and are convinced of the manager’s ability to effectively manage human resource management. In this case, HRM is likely to be more strongly related to PO-fit. Furthermore, employees with high trust in the manager are likely to experience strong commitment from their manager. Therefore, employees tend to be strongly affiliated with and loyal to their organization. Based on the norm of reciprocity, employees feel obligated to care about the organization’s welfare and to help the organization reach its objectives. Employees’ willingness to maintain membership with the organization because enhanced trust in the manager will lower employees intentions to leave the organization.
In contrast, employees with low trust in the manager are not convinced about the ability, benevolence and integrity of their manager. They do not believe that their manager have the ability to effectively manage human resource management. They do not have the feeling that their manager values their contribution and cares about the welfare of the employee. Finally, they have the feeling that the values of the manager and not in line with their own values. Thus, when trust in the manager is low, HRM is likely to be less strongly related to PO-fit. Employees are less connected and loyal to the organization and are less willing to maintain membership with the organization. Employees with low trust in the manager are expected to have higher turnover intentions (Mayer et al., 1995).
Overall, if there is congruence between the values of an employee and of the
organization, person-organization fit is achieved. This can be achieved with a consistent set of human resource practices through which the manager can communicate the most important values. The manager is responsible to communicate the HR practices and policy. Therefore, it is expected that the relationship between HRM and PO-fit and in turn employee turnover intentions depends on the level of trust in the manager. Open, transparent and consistent communication helps employees to develop confidence in their manager. In this way,
employees feel strongly committed to the organization and therefore these employees are less likely to leave the organization. Thus they are more likely to respond positively to HRM compared to a situation in which low trust is perceived. Employees who have low trust in their manager respond more negatively to HRM, because they do not have the feeling that the values of the manager communicated through HRM are in line with their own values.
Employees are less connected and loyal to the organization and are less willing to maintain membership with the organization. Thus, a mediated moderation model is proposed, which indicates that the indirect relationship between HRM and turnover intentions through PO-fit is stronger for employees with high trust in the manager than for employees with low trust in the manager.
Hypothesis 3: Trust in the manager moderates the relationship between HRM on PO-fit and in turn turnover intentions, (such that the mediated relationship will be weaker for low trust in the manager than for high trust in the manager).
3. Method
3.1 Procedure and sample description
This study was conducted by a research team consisting of six students of the Master Business Studies. The study was supervised by Dr. Deanne N. den Hartog and Dr. Corine Boon from the Human Resource Management – Organizational Behavior department of the University of Amsterdam. The aim was to collect data for potential academic papers for the professors and for the Master theses of the students. Another aim of the research team was to obtain data from both manager and subordinate (employee) perspectives. Therefore, responses were collected in sets of one manager with one to three associated subordinates. The employee and manager surveys were linked using a matching code and the data was online collected via surveymonkey.com.
So, a quantitative multi-source study was conducted with four internet-mediated questionnaires; an English employee and manager questionnaire and a Dutch employee and manager questionnaire. Translation procedures were followed to convert original English items to Dutch by bilingual researchers from the department. The employee questionnaire took approximately 20 minutes to complete. The manager questionnaire took approximately 10 minutes to complete. The questionnaires consisted of items to test a wide range of
variables like; commitment, trust, communication, engagement, HRM, personality traits, job characteristics, leadership styles etc. The current study only uses the data obtained for the variables HRM, PO-fit, employee turnover intentions and trust in the manager. Managers rated HRM and employees were asked to rate PO-fit, turnover intentions and trust in the manager.
This research built on the data that is obtained by previous year’s research team 2011-2012. Last year’s research team collected responses from February to April 2011-2012. The questionnaires were available in Dutch and in German. The current year’s research team collected responses from March to May 2013. The only change made to the previous year’s survey pertinent to this study was the addition of average hours working a week and country as control variables.
The current study used a combination of non-probability sampling methods, namely convenience and snowball sampling. Convenience sampling was used to gain access to organizations and respondents through personal contacts from the research team. In addition, known organizations were invited via e-mail, using contact information from their corporate
websites. The purpose was to create a heterogeneous sample representing general employee and manager attitudes and behaviors. Respondents were recruited from several countries, industries, organizational levels, and from non-profit, public, and private sectors. Respondents in Dutch organizations received Dutch questionnaires, and respondents in other countries received an English version of the questionnaire. Complementary, snowball sampling was used through already participating managers and employees, who were trying to get more managers and employees within the organization involved (Bryman, 2012). All respondents were invited via e-mail with a formal letter. First, emails were sent to ask for consent and to give instructions on the procedure of the online questionnaire. Second, the email addresses of the participating manager and subordinates were asked to send special links to each
respondent. Reminders were sent after two weeks to the respondents who had not yet finished the questionnaire. Respondents were able to contact the research team for questions by email via uva.hrmresearchgroup@gmail.com.
The questionnaire contained a clear introduction where the purpose, the duration, and the different concepts of the questionnaire were outlined. It emphasized that information obtained from the surveys will be treated confidentially and will not be used for other purposes than this study. As a thank you for their time and effort, there was a possibility to indicate whether the respondent would like to receive an overall research summary of the six Master’s students’ studies.
The sample of preceding year’s research team consisted of 56 managers and 174 employees. The sample of this year’s research team consisted of 86 managers and 197
employees. The two datasets were merged and the final sample consisted of 371 usable dyads of managers and corresponding employees. Missing data reduced the sample from 415 to 371 usable dyads. The combined sample size therefore consisted of 142 managers and 371
employees, which is in total 513 respondents. Respondents worked in a variety of industries including education and research, banking, sales, health care, manufacturing, information technology (IT), and natural resources. Data was mainly collected from respondents working in the Netherlands (61%), but also included samples from Germany (12.7%), Romania (10.9%), Greece (7.5%), and the United States (5.5%). There were small samples from Canada, France, Italy, and Spain.
This year’s research team approached 103 managers and 247 employees, from which 96 managers and 226 employees responded. The response rates were 90,57% for the
92%. These response rates are significantly high and this can be explained by the fact that almost all the respondents are personal contacts from the researchers. This makes it more plausible that many of the respondents who were contacted also responded. The final sample included 197 (53%) female employees and 174 (47%) male employees. These gender
numbers are closer than those of the manager responses, which included 92 (65%) male managers and only 50 (35%) female managers. The average age of employees was 35 (SD = 11.55). The average age of managers was 43 years (SD = 12.69). The employee and manager respondents were between 18 and 62 years old. Responses showed that 57% of managers had a university bachelor degree or higher compared to 41% of employees. The average
educational background of employees was higher vocational education and they worked on average 36 hours a week. Employees worked on average for 7.91 years (SD = 8.62) with their current organization, and with their current managers for 2.45 years (SD = 2.83).
3.2 Description of measures
This section shows the measurement scales of all the variables included in this study. All items were rated on a 7-point Likert scale ranging from 1 (strongly disagree) to 7 (strongly agree).
Employees completed the 3-item turnover intentions scale developed by Colarelli (1984). Sample items are: “If I have my own way, I will be working for this organization one year from now”, “I frequently think of quitting my job” and the last item is “I am planning to search for a new job during the next 12 months”. Cronbach’s alpha was .78.
HRM is measured by 15 items (Kehoe and Wright, 2013). Managers were asked to assess to what extent they agreed with specific statements related to specific practices. An HRM scale should contain the three components; ability enhancing, motivation-enhancing and opportunity-enhancing practices. The HRM scale measured in this study includes all the practices. The 15 items together reflect one high performance human resource system. These items measure manager perceptions of high performance human resource practices. Sample items are: “Employees receive formal training each year”, “Employees take formal pre-employment tests (paper and pencil or work sample)” and “At least once a year, employees receive a formal evaluation of their performance”. Cronbach’s alpha was .80.
PO-fit defined as value congruence is measured by 3 items (Cable & DeRue, 2002). Sample items are: “The things that I value in life are very similar to the things that my organization values”, “My personal values match my organization’s values and culture” and
“My organization’s values and culture provide a good fit with the things that I value in life”. Cronbach’s alpha was .91.
There are 13 items measuring trust in your manager (Mayer & Davis, 1999). Sample items are: “I never have to worry about whether my manager will stick to his/her word”, “My manager will go out of his/her way to help me” and “I feel very confident about my
manager’s skills”. Cronbach’s alpha was .82.
3.3 Analytical strategy
In this section the analytical strategy will be discussed and in the next section the results of these analyses will be presented. The relationships between variables were tested through an explanatory deductive research approach.
First, the data was imported and analyzed by using Statistical Package for the Social Sciences 17.0 (SPSS 17.0). Only the respondents who skipped the majority of the questions were excluded from the dataset. To deal with missing data, missing values were marked by 999. These missing values were replaced by the use of HOTDECK imputation (Myers, 2011). Imputation is the process of replacing missing data with substituted values. The dataset was checked for outliers with stem-and-leaf plots. Furthermore, there was one counterbalanced item that had to be reversed coded. It is one of the items of employee turnover intentions; “If I have my own way, I will be working for this organization one year from now”. The rest of the items were positively formulated. The internal consistency of the scales consisted of different items was tested through reliability tests. According to Field (2009), scales are reliable when the Cronbach’s Alpha is higher than 0.7. Following, the scales were computed by averaging the items. Subsequently, exploratory analyses were computed to get demographic
characteristics of the sample. A dummy variable was made from the control variable country, in which 1 refers to respondents from the Netherlands and 0 refers to respondents from other countries. Furthermore, to explore the strength of the relationships between the scales
bivariate correlations analyses were done. Preliminary analyses were conducted before regression analyses to test assumptions of normality and multicollinearity. Finally, the hypotheses were tested through multiple linear regression analyses. Specifically, the mediation effects were investigated by four steps according to the method of Baron and Kenny (1986). The overall moderated mediation model is tested by a SPSS macro called MODMED provided by the website of Andrew Hayes (www.afhayes.com). This syntax, once executed, creates a new command in SPSS called MODMED and conduct analyses of
conditional indirect effects, commonly known as mediated moderation effects. This study used example model 2, in which some fourth variable (W/moderator) affects the a1 path between the independent variable (X) and the mediator (M). The MODMED command is used to give information about which variables in the model functions as the independent variable, the mediator, the outcome, the moderated and control variables in the model. The independent and moderator variables are mean centered and multiplying the centered variables creates the interaction term, following suggestions of Preacher et al. (2007).This mediated moderation model attempt to explain both how and when a given effect occurs (Preacher et al., 2007). In other words, mediated moderation analyses identifies whether employees high and low on trust in the manager benefit differently from HRM (Donaldson, 2001; Aiken & West, 1997).
Figure 3. Mediated moderation model (Preacher et al., 2007).
In order to prevent omitted-variable bias, six controls were included in the analyses: age, gender, time working for the current organization, time working for the current manager, average hours working a week and country of the employee. Employee country is assumed to also be the country where the corresponding manager is living. Country is included as a control variable, because of the impact that cultural differences have on HR practices and policies and its effectiveness (Laurent, 1986). Only employee controls are included in the study, because only the independent variable HRM is rated by the manager of which is expected to influence all the other variables in the model rated by employees.
4. Results
4.1 Data preparation
Table 2 presents means, standard deviations, correlations and Cronbach’s alphas of the variables. It shows that all variables have a Cronbach’s alpha higher than 0.7, which means that all variables are reliable (mHRM α=8.12, ePO-fit α=.957, eTrustMan α=.941, eTurnInt α=.807). No items were deleted in order to increase the internal consistency. HRM and trust in the manager are positively related to PO-fit (r=.151** and r= .379**). Besides, HRM, trust in the manager and PO-fit are significant negatively related to turnover intentions (respectively; r= -.115*, r=-.244**, r=-.427**).
To check for normality a Kolmogorov-Smirnov test is done. Table 1 presents the skewness, kurtosis and values of the Kolmogorov-Smirnov test. The skewness presents ‘the extent to which values of a distribution deviates from symmetry around the mean’ and the kurtosis refers to ‘the peak or flatness of a distribution’ (Field, 2009, p. 413). In order to find a normal distribution, these values should be equal or approximately zero.
Table 1. Normality test.
Skewness Kurtosis Statistics Sig.
1. mHRM .13 .26 .09 .00
2. ePO-fit .13 .26 .13 .00
3. eTrustMan .13 .26 .15 .00
4. eTurnInt .13 .26 .09 .00
Table 2. Descriptives, correlations, and reliabilities of the study variables.
Mean SD 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. Controls 1. eGender 2. eAge 35.30 11.55 -.195** 3. eTimeOrg 7.91 8.62 -.209** .725** 4. eTimeMan 2.45 2.83 -.098 .405** .432** 5. eAvgHrWk 35.96 11.14 -.382** .334** .189** .126 6. eCountry Variables 7. mHRM 4.87 .82 -.076 -.059 -.047 -.174** .114 .098 (.812) 8. ePO-fit 5.02 1.26 .139** -.035 -.043 .017 -.125 .221** .151** (.957) 9. eTrustMan 2.54 1.46 .056 -.101 -.145** -.038 -.180* .114* .148** .379** (.941) 10. eTurnInt 5.53 .92 -.091 -.049 -.012 -.016 .147* -.126* -.115* -.427** -.244** (.807) Notes: Alphas are in parentheses; *p < .05. **p < .01.
The output of the Kolmogorov-Smirnov test shows that the values of all the variables are significant. This means that the scores on the variables are not normal distributed. The
Kolmogorov-Smirnov test rejects the assumption of normality. All the variables were skewed to the right, except from turnover intentions, which was skewed to the left. These skewed variables were not corrected. A possible explanation of the skewed variables is that three of the four variables, namely PO-fit, trust in the manager and turnover intentions, were self-rated by the employee. The hazard of self-rated items are for example social desirable answers, which can lead to scores above the mean (for turnover intentions below the mean).
Participation was not fully anonymous, because special codes were made to match managers and subordinates. Employees could be afraid of losing their job if they honestly admitted that they are thinking about quitting the job. Complementary, employees could be afraid to said something negative about their manager, for example that they do not trust their manager. Tests for multicollinearity demonstrated that there are no multicollinearity problems in any of the regression analyses, because the tolerance levels are higher than 0.10, and the VIF values are below 5.
Table 3. Multicollinearity analysis.
Tolerance VIF
mHRM .969 1.032
ePOfit .851 1.176
eTrustMan .850 1.176
Notes: Tolerance levels needs to be higher than .10 and VIF values below 5. 4.2 Analyses
To test the hypotheses regression analyses were conducted and all regressions are controlled for employee gender, age, years working by the current organization, years working for the current manager, average hours of working a week and country.
Hypothesis 1: HRM is negatively related to employee turnover intentions.
To test the first hypothesis, a direct relationship between HRM and turnover intentions, a linear regression analysis is done. The results show that the regression model is significant (F= 4.67, p= .03). There is a direct significant negative relationship between HRM and turnover intentions (β= -.20, t=-2.16, p=.03). The adjusted R² is 0.8%.
Hypothesis 2: PO-fit mediated the relationship between HRM and turnover intentions. The mediation effect of PO-fit between HRM and turnover intentions is tested through four regression analyses (Baron & Kenny, 1986). The first regression analysis test if there is a
direct positively relationship between HRM and turnover intentions (see hypothesis 1). There is a direct significant negative relationship between HRM and turnover intentions (β= -.20, t=-2.16, p=.03). The second regression analysis test if there is a direct positive relationship between HRM and PO-fit. The regression model as a whole is significant (F= 2.98, p= .01). The control variables and HRM explain 10 % of the variance of PO-fit (R²= .100; Adjusted R²= .066). The relationship is positively significant (β= .35, t= 3.02, p=.00). The third regression analysis test if there is a direct positive relationship between PO-fit and turnover intentions. The regression model is significant (F=8.35, p=.00). Results show that PO-fit is significantly related to turnover intentions (β = -.44, t= -6.98, p= .00).The control variables and PO-fit explain 24.2 % of the variance in turnover intentions (R²=.242; Adjusted R²=.213). The last regression analysis test if the relationship between HRM and turnover intentions is mediated by PO-fit. The regression model as a whole is significant (F= 7.22, p=.00). PO-fit is a full mediator between HRM and turnover intentions, because PO-fit is significant (β=-.45, t=-6.85, p=.00) and HRM is not significant anymore (β=.04, t=3.75, p=.71). The controls, HRM and PO-fit explain 24.3% of the variance in turnover intentions (Adjusted R²= .209). The outputs of the multiple regressions are located in table 5. Hypothesis 2 is supported.
Table 4. Multiple regressions: HRM and turnover intentions mediated by PO-fit.
Variables Step F p Adj R² β t p eGen -.239 -1.13 .26 eAge -.008 -.634 .53 eTimeOrg .017 1.138 .26 eTimeMan -.028 -.821 .41 eAvgHrWk .014 1.434 .15 eCountry .014 .065 .95 HRM - TurnInt 1 4.67 .03* .008 -.20 -2.16 .03* eGen .127 .613 .541 eAge -.011 -.897 .371 eTimeOrg -.002 -.166 .868 eTimeMan .019 .541 .589 eAvgHrWk -.013 -1.281 .202 eCountry -.246 -1.204 .230 HRM – PO-fit 2 2.98 .01** .066 .35 3.02 .00** eGen -.176 -.937 .35 eAge -.013 -1.207 .23 eTimeOrg .015 1.124 .26 eTimeMan -.026 -.836 .40 eAvgHrWk .010 1.114 .27 eCountry -.061 -.331 .741 PO-fit – TurnInt 3 8.35 .00** .213 -.44 -6.98 .00** eGen -.192 -1.020 .31 eAge -.012 -1.112 .27 eTimeOrg .015 1.096 .28 eTimeMan -.022 -.709 .48 eAvgHrWk .008 .909 .37 eCountry -.084 -.457 .65 HRM + PO-fit – TurnInt 4 7.22 .00** .209 HRM .04 3.75 .71 PO-fit -.45 -6.85 .00** Notes: *p < .05. **p < .01.
Hypothesis 3: Trust in the manager moderates the relationship between HRM on PO-fit and in turn turnover intentions, such that the mediated relationship will be stronger when employees perceive trust in the manager as high than employees perceive trust in the manager as low.
Hypothesis three proposes a mediated moderation model. Mediated moderation refers to an effect of the independent variable on the dependent variable via a mediator variable, which is different depending on the levels of a moderator variable(Preacher et al., 2007). Specifically, in this research the effect of HRM on turnover intentions via PO-fit depends on the level of trust in the manager. A mediated moderation is also known as a conditional indirect effect.
Conventionally, moderation and mediation were tested separately, which leads to various methodological problems. For example, the accuracy and utility is undermined (Edwards & Lambert, 2007; Preacher et al., 2007). Therefore, we tested mediation and moderation simultaneously using regression analyses with a syntax provided by Preacher and Hayes. PO-fit was regressed on the control variables, high-performance human resource practices, trust in the manager and the interaction term (HRM x trust in the manager). The first column of table 6, with PO-fit as the dependent variable (mediator model),shows that HRM is significant related to PO-fit (β=.19, p<.01) and trust in the manager is also significant related to PO-fit (β=.43, p<.01). Furthermore, the interaction term HRM x TrustMan is significant (β = -.14, p <.05), which indicates that trust in the manager function as a moderator in the relationship between HRM and PO-fit.
Figure 4. The moderating role of trust in the manager between HRM and PO-fit.
Table 5. Results of hierarchical regression analysis.
PO-fit Turnover Intentions
Gen -.01 -.15 Age -.00 -.01 TimeOrg .00 .01 TimeMan -.03 -.00 AvgHrWk -.01 .01 Country .00** -.00 HRM .19** .05 TrustMan .43** -.12 HRM x TrustMan .14* .08 PO-fit -.32**
Notes. Standardized coefficients are presented; *p < .05. **p < .01.
The second column of Table 6 , with turnover intentions as the dependent variable
(=dependent model), shows that PO-fit is significant (β = -.32**, p<.01). This indicates that PO-fit serves as a mediator between HRMxTrust and turnover intentions.
The strength of the indirect effect was tested for specific values of the moderator. These specific moderator values are the sample mean, one standard deviation below and above the sample mean of trust in the manager. Results show a significant positive
relationship for high trust in the manager, and a nonsignificant relationship for low trust in the manager. The indirect effect of HRM on turnover intentions via PO-fit was significantly
1 2 3 4 5 6 7 Low HRM High HRM PO -f it Low TrustMan High TrustMan
negative for high (+1SD: B = -.11, p=.01) and mean scores on trust in the manager (B = -.06, p = .02), whereas no significant indirect effect is found for low trust in the manager (-1SD: B =-.02, p = .57). These results provide support for hypothesis 3.
5. Discussion
5.1 Conclusion
The aim of this study was to examine potential mechanisms through which HRM affect employees turnover intentions. The current study investigated the mediating effect of PO-fit and the moderating effect of trust in the manager. In other words, the current study
investigated how and when HRM is related to employee turnover intentions. The results from 371 usable manager-employee dyads revealed that HRM is negative related to employee turnover intentions, which supports hypothesis 1. The results of testing the mediating effect of fit on the relationship between HRM and employee turnover intentions, revealed that PO-fit had a fully mediated effect, which supports hypothesis 2. Finally, trust in the manager moderates the relationship between HRM on PO-fit and in turn turnover intentions, such that the mediated relationship will be stronger when employees perceive trust in the manager as high than employees perceive trust in the manager as low. Hereby, hypothesis 3 is also supported. To conclude, HRM not only had a direct effect, but also an indirect effect on employee turnover intentions through their influence on PO-fit and trust in the manager. Thus, HRM, PO-fit and trust in the manager are all three important independent factors explaining turnover intentions of employees.
5.2 Theoretical and managerial implications
This study contributes to the SHRM literature by opening the black box between HRM and employee performance. The mechanisms through which HPWPs are linked to employee outcomes have received little attention from researchers (Boon et al., 2011; Mostafa & Gould-Williams, 2014). This quantitative multi-sourced study has addressed this research gap by examining the mediating effect of PO-fit and the moderating effect of trust in the manager, on the relationship between HRM and employees turnover intentions. High performance work practices, constructed as integrated and coherent systems of mutually reinforcing practices, enhance the three main components of the AMO theory; ability, motivation and opportunities to participate for employees. This affect the attitudes and behaviors of employees in such a way that they are less likely to leave the organization. HPWPs designed and implemented according to the AMO theory will result in employees performing behaviors consistent with organizational goals and strategy (Boselie, 2010; Kehoe & Wright, 2013).
Further, based on the attraction, selection and attrition theory, it can be concluded that Person-Organization fit fully mediates the relationship between HRM and turnover intentions.
