绩效管理文献3

Putting employee involvement in context: a cross-level model examining job satisfaction and absenteeism in high-involvement work systems Christopher D. Zatzick* and Roderick D. Iverson Faculty of Business Administration, Simon Fraser University, Burnaby, BC, Canada The current study examines how high-involvement work systems (HIWS) influence employee responses to involvement initiatives. While existing research has linked HIWS to individual attitudes, we predict that an organization’s HIWS moderate the relationship between employee involvement and job satisfaction and absenteeism. Using multilevel data (8454 employees from 1429 workplaces), we found that employee involvement and HIWS are positively related to employee job satisfaction. Additionally, the results support a cross-level interaction: at high levels of HIWS, employee involvement is negatively related to absenteeism, whereas at low levels of HIWS, the negative relationship is weaker. The implications of the findings are discussed. Keywords: absenteeism; employee involvement; high-involvement work systems; multilevel research; organizational context Over the past two decades, high-involvement work systems (HIWS) have received significant attention in the literature. HIWS consist of synergistic and mutually reinforcing human resource (HR) practices that are designed to increase employees’ ability, motivation, and opportunity to contribute to the organization, which in turn is thought to positively impact performance. Many studies report positive relationships between HIWS (also referred to as high-performance or high-commitment work systems) and performance at the organization or establishment levels (Arthur 1994; Huselid 1995; Guthrie 2001; Datta, Guthrie and Wright 2005). A recent meta-analysis of 92 studies (Combs, Liu, Hall and Ketchen 2006) found a positive relationship between the bundles or systems of HR practices and the various performance outcomes of productivity, accounting returns, growth, and market returns. Hence, researchers agree that HIWS improve firm outcomes, in part, through their ability to develop, motivate, and retain human capital. Despite the consensus that HIWS influence performance through employees’ attitudes and behaviors, a limited amount of empirical research exists on this issue (Wright and Boswell 2002; Butts, Vandenberg, DeJoy, Schaffer and Wilson 2009). In particular, few studies have examined cross-level relationships between HIWS and employee attitudes and behaviors (see Takeuchi, Chen and Takeuchi 2009 for a recent exception). Rather, studies have focused on employee reports of HR practices and how these self-reports relate to attitudes such as job satisfaction and commitment (e.g. Macky and Boxall 2007; Mohr and Zoghi 2008). Employees who report greater levels of involvement (e.g. working in teams, participating in upward feedback programs, and receiving information about ISSN 0958-5192 print/ISSN 1466-4399 online q 2011 Taylor & Francis http://dx.doi.org/10.1080/09585192.2011.561016 http://www.tandfonline.com *Corresponding author. Email: czatzick@sfu.ca The International Journal of Human Resource Management, Vol. 22, No. 17, October 2011, 3462–3476 organizational results) are generally more satisfied and committed to the workplace. However, by focusing solely on individual perceptions of involvement, these studies have to a large extent ignored the role of the overall HR system in this relationship. Thus, more research is needed to understand the complex interplay between organization-level practices, employee involvement, and the resulting attitudes and behaviors. The current study addresses an important gap in the literature by examining both individual reports of involvement and the context in which the involvement is provided. Specifically, we focus on HIWS as an important contextual factor that conveys the value and importance of employee involvement to the organization (Bowen and Ostroff 2004). Drawing on Johns’ (2006) multilevel framework, we argue that HIWS shape the meaning employees ascribe to the involvement practices and provide the support and resources necessary for employees to utilize their involvement more effectively. We test our predictions using a multilevel data set consisting of 1429 organizations and 8424 employees. Theory and hypotheses The importance of involvement for employees and organizations is well established in the literature. The concept of employee involvement is typically applied to lower level employees and focuses on their opportunities to make decisions about their jobs, as well as their participation in the whole business. Perhaps the most prominent stream of research related to employee involvement is the job characteristics model (JCM; Hackman and Oldham 1980). The JCM suggests that organizations can design jobs to increase employee involvement by enriching jobs with more autonomy and decision-making authority. Job enrichment provides employees the opportunity to have meaning and responsibility in their work. Researchers have found that job satisfaction is increased when jobs are enriched with more involvement and discretion (Fried and Ferris 1987) and greater skills and autonomy (Berg 1999).1 Finally, Humphrey, Nahrgang and Morgeson (2007) in their meta-analysis of 259 studies reported the five core characteristics of the JCM to have corrected correlations ranging from 0.31 to 0.48 with job satisfaction. In addition to influencing job satisfaction, involvement is thought to be an important motivator for employees to attend work. Employee absenteeism, which refers to the temporary withdrawal from the organization, is used by employees to escape from a work situation considered as stressful and unpleasant (Deery, Iverson and Walsh 2006). When employees experience job enrichment and involvement through greater autonomy and responsibility in their work, they will be less likely to withdraw. In Harrison and Martocchio’s (1998) review of the absenteeism literature, they report evidence of a moderate negative relationship between job characteristics (i.e. increased autonomy and involvement) and absenteeism. These findings are also supported in the recent meta-analysis by Humphrey et al. (2007). Hence, increased involvement via job enrichment and voice is valued by employees and leads to positive attitudes and behaviors. Thus, we hypothesize Hypothesis 1: Employee involvement is positively related to job satisfaction. Hypothesis 2: Employee involvement is negatively related to absenteeism. High-involvement work systems Lawler (1992) extended the focus beyond individual-level involvement by suggesting that organizations can build HIWS in order to improve overall organizational performance. Lawler (1992) identified four ways that HIWS influence employee involvement and The International Journal of Human Resource Management 3463 engagement in the workplace including power, information, rewards, and knowledge. First, by giving employees the power to act, organizations allow individuals to experience greater autonomy and control in their job. An example of an HR practice that increases employee power is the use of self-directed work teams, which increase employee decision- making authority over job-related activities such as planning tasks, staffing, and follow-up of results. Second, if an organization wants employees to make sound decisions, it must provide information about key processes, events, and results. Information sharing practices include company newsletters and upward feedback mechanisms (such as employee surveys and suggestion programs). Third, to ensure that employee motivation is consistent with an organization’s goals, the organization must provide rewards that are linked to group outcomes, as well as individual capabilities and contributions. Compensation practices, such as gainsharing and profit sharing, are examples of how organizations can align employees’ behavior with the organization’s goals. Finally, Lawler (1992) emphasized the importance of ensuring that employees have knowledge of the whole work systems. Work design practices such as job rotation and cross-functional teams seek to enhance employees’ understanding of the larger organizational systems. In addition, training and development opportunities provide employees with the knowledge and skills needed to succeed in the organizations. Research (i.e. Vandenberg, Richardson and Eastman 1999; Butts et al. 2009) has supported the link between HIWS and the extent to which employees report having power, information, rewards, and knowledge. Furthermore, a number of studies have found that HIWS (based on the HR practices and involvement opportunities available) are positively related to job satisfaction and commitment (Vandenberg et al. 1999; Harley, Allen and Sargent 2007; Macky and Boxall 2007; Mohr and Zoghi 2008). Recently, Takeuchi et al. (2009) found support for a cross-level relationship where employees were more satisfied and committed when working in a high-performance work system, in part, from the climate created by the synergistic and mutually reinforcing HR practices. Another potential benefit of HIWS is that employees will be more motivated to attend work. Several studies have linked an organization’s system of HR practices to unit-level absenteeism. Richardson and Vandenberg (2005) found that work units with a strong climate of involvement have lower levels of absenteeism. Current research by Guthrie, Flood, Liu and MacCurtain (2009) also reported that an organization’s high-performance work system was negatively related to organization-level absenteeism. Building on these findings, we expect that HIWS will have a cross-level relationship with individual absenteeism. Hence, we make the following predictions: Hypothesis 3: HIWS are positively related to employee satisfaction. Hypothesis 4: HIWS are negatively related to employee absenteeism. In addition to having a direct effect on employee attitudes and behaviors, HIWS also serve as a contextual factor that influences how employees respond to their involvement. In the following section, we examine the intersection of HIWS and employee involvement with respect to individual job satisfaction and absenteeism. HIWS as a contextual factor To examine HIWS as a contextual factor, we draw on Johns’ (2006) multilevel framework that identifies two primary dimensions of organizational context: omnibus context and discrete context. The omnibus context consists of global aspects of the context, such as a sample of employees in a particular occupation, location, or point in time, whereas the C.D. Zatzick and R.D. Iverson3464 discrete context consists of factors directly related to a specific situation such as the task, social, and physical environments. Given the nature of our sample (large, multi-industry, geographically dispersed, and multilevel), we focus on the discrete context as the task and social environments are closely related to HIWS. Specifically, four factors of the discrete context identified by Johns’ (2006) including uncertainty, autonomy, accountability, and the norms associated with the social context are influenced by HIWS.2 First, the degree of uncertainty in an organizational context shapes the meaning participants ascribe to the situation (Johns 2006). In order for HIWS to succeed, it must elicit unambiguous and shared perceptions about the strength and purpose of the system (Bowen and Ostroff 2004). In other words, HIWS affect how employees make sense of their involvement in the organization. When more elements of HIWS are present it conveys a clear and consistent message about management’s intent to develop a commitment-oriented HR system, which will reduce the uncertainty regarding the HIWS. In contrast, the presence of relatively few HR practices reflects low levels of HIWS and sends an inconsistent message about the value of employees, which increases employee uncertainty about their involvement. Johns (1999) discusses how high levels of uncertainty in an organizational context increase employee self-serving behavior such as absenteeism. When fewer elements of HIWS are present in an organization, we expect the uncertainty about the organization’s HR system to be greater, and the benefits of employee involvement to be lower. Second, Johns (2006) identified the autonomy provided by the organizational context as an important influence on employee attitudes and behaviors. At the individual level, employees respond positively to HR practices designed to increase autonomy (Hackman and Oldham 1980). Again, however, employees may view involvement practices that increase autonomy as insufficient if they are not supported by other practices that align with their autonomy. This is consistent with the idea that employees are receiving ‘false empowerment’ (Argyris 1998) where their involvement is superficial; a token offering from management without any real power to participate. We posit that HIWS provide the foundation for employees to have decision-making authority and enable them to benefit from receiving greater involvement in the organization. In sum, employees will not experience the desired autonomy when working in an organization with low levels of HIWS, even when receiving some token involvement. Third, Johns (2006) identifies the level of accountability in an organizational context as a major influence on employee behavior. As discussed above, increasing employee involvement will increase accountability in the sense that employees are responsible for their decisions and outcomes. For example, group-level compensation such as gainsharing is designed to increase participation in a team environment and to focus on effort toward organizational goals (Lawler 1992). Hence, the level of accountability is likely greater in organizations with higher levels of HIWS. Finally, an organization’s social context will influence employee attitudes and behaviors (Johns 2006). In organizations with high levels of HIWS, managers and employees will utilize their discretion in a more open and transparent manner, which will help employees learn about exercising their decision-making authority including when and how to do it. Hence, we expect employee responses to involvement to be more positive in HIWS due to the norms and expectations present in the system. To summarize, we posit that the relationship between employee involvement and satisfaction and absenteeism will be moderated by an organization’s HIWS (see Figure 1). An organization with high levels of HIWS has adopted a breadth of HR practices designed to increase employees’ power, information, knowledge, and rewards. High levels of HIWS The International Journal of Human Resource Management 3465 convey the value and importance of employee participation to the organization, whereby employees will respond more positively about their involvement and experience greater satisfaction and lower absenteeism. In contrast, employees in an organization with low levels of HIWS perceive less support and view their participation as ‘false empowerment’ that enlarges their job rather than enriches it. Thus, we predict Hypothesis 5: The relationship between employee involvement and job satisfaction is moderated by HIWS. Specifically, the positive relationship will be stronger for employees in organizations with high levels of HIWS as compared to employees in organizations with low levels of HIWS. Hypothesis 6: The relationship between employee involvement and absenteeism is moderated by HIWS. Specifically, the negative relationship will be stronger for employees in organizations with high levels of HIWS as compared to employees in organizations with low levels of HIWS. Method Participants and design To test our hypothesis, we use the Workplace and Employee Survey (WES) conducted by Statistics Canada. The survey randomly sampled business establishments and a subset of employees within those establishments with up to 20 employees per location surveyed. Because these are stand-alone business locations with a discrete workforce and HIWS, we refer to these as ‘organizations’ going forward. Response rates in 1999 and 2000 for the sample of organizations were 95.2 and 85.9%, respectively. Respondents of the organization survey included senior managers, general managers, and owners, depending on the size of the organization. In many cases, multiple responses were used to ensure that the information provided was accurate. A complete description of the data collection procedures can be found in the WES compendium (Statistics Canada 2004). Once organizations agreed to participate, employees were randomly selected from a list of employees on the payroll for the specific location. Because of our focus on HIWS, we dropped organizations that had 10 or fewer employees because the sophistication of the HR system is limited. In addition, we conducted listwise deletion for missing data. Finally, consistent with recent hierarchical linear modeling (HLM) studies (Bacharach, Bamberger High-Involvement Work System Employee Involvement Level 2 Level 1 Job Satisfaction Absenteeism Figure 1. Multilevel model of HIWS, employee involvement, and individual outcomes. C.D. Zatzick and R.D. Iverson3466 and Doveh 2008; Giardini and Frese 2008), we restricted our sample to organizations with at least five employee respondents. Thus, our final sample included 8454 employees from 1429 organizations. Analyses revealed no significant differences in HIWS between organizations with five or more respondents and those which were dropped from having less than five respondents. Finally, we compared organizations present in both years to those missing in 2000 and found no differences in HIWS and size (t , 1 and p . 0.1). Measures Absenteeism We measure absenteeism with information taken from the 2000 employee survey. We use self-reported information on the number of paid sick days taken in the previous 12 months. One year is considered as an appropriate period of time to capture a trend in absenteeism, particularly when linking it to job characteristics and job satisfaction (Harrison and Martocchio 1998). In the current study, absenteeism was approximately 7 days per employee. Because of the skewness of the variable (Hammer and Landau 1981), we use the logarithm of paid sick days. Job satisfaction Job satisfaction is measured from the 2000 employee survey using two items (1 ¼ very dissatisfied to 4 ¼ very satisfied). Item 1 focuses on global satisfaction: ‘Considering all aspects of this job, how satisfied are you with the job?’ Item 2 focuses on pay satisfaction: ‘How satisfied are you with your pay?’ Although the two items can be problematic for certain constructs, researchers have demonstrated the reliability of single item measures of job satisfaction (Wanous, Reichers and Hudy 1997). The coefficient alpha was 0.65, which is reasonable for measures with only two items (Cortina 1993). Employee involvement Employee involvement was measured from the 1999 employee survey. We focused on HR practices that can increase employee involvement via increased power, information, rewards, and knowledge in their jobs (Lawler 1992). Employees were asked, on a three- point scale (1 ¼ never, 2 ¼ occasionally, and 3 ¼ frequently), how often they participated in or received the following: (1) employee surveys, (2) employee suggestion programs, (3) job rotation or cross-training programs, (4) labor–management committees, (5) quality circles, (6) self-directed work groups, and (7) newsletter with information about workplace performance, organizational changes, or the implementation of new technology. In addition, we added an eighth ite