Organization Development

The Journal of Applied Behavioral Science 49(4) 413 –436

© 2013 NTL Institute Reprints and permissions:

sagepub.com/journalsPermissions.nav DOI: 10.1177/0021886313479998

jabs.sagepub.com

Article

How High-Involvement Work Processes Increase Organization Performance: The Role of Organizational Citizenship Behavior

Mark A. Kizilos1, Chailin Cummings2 and Thomas G. Cummings3

Abstract Employee involvement is a popular approach to improve organization performance. It moves decision making downward in the organization so employees can make decisions and solve problems as quickly and close to their source as possible. One of the most developed and referenced approaches to involvement is Edward E. Lawler’s model of “high-involvement work processes” (HIWP). It describes organizational attributes that contribute to employee involvement and explains how they work together to increase organization performance. Although extensive attention has been paid to Lawler’s model in the literature, empirical tests of the model are still in a preliminary stage. Our study describes and tests a mechanism through which HIWP increases organization performance, organizational citizenship behavior. We find that organizational citizenship behavior mediates the relationship between HIWP and organization performance in a sample of 143 consumer-products organization units. Results also confirm that the HIWP attributes work together synergistically to create opportunities for employee involvement.

Keywords employee involvement, high-involvement work processes, organizational citizenship behavior, organization performance

1Experienced-Based Development Associates, MN, USA 2California State University, Long Beach, CA, USA 3University of Southern California, Los Angeles, CA, USA

Corresponding Author: Thomas G. Cummings, Marshall School of Business, University of Southern California, Los Angeles, CA 90089-0808, USA. Email: tcummings@marshall.usc.edu

479998 JAB49410.1177/0021886313479998The Journal of Applied Behavioral ScienceKizilos et al. research-article2013

at University of Maryland Baltimore Health Sci & Hum Serv Lib on September 27, 2016jab.sagepub.comDownloaded from

414 The Journal of Applied Behavioral Science 49(4)

Faced with competitive pressures to operate leaner, more efficiently and flexibly, organizations are increasingly applying employee involvement to achieve greater performance and competitive advantage (cf. Lawler & Boudreau, 2009). Employee involvement empowers employees to make decisions and solve problems. It moves decision making downward in the organization to the lowest level where a compe- tent decision can be made (Conger & Kanungo, 1988; Lawler, Ledford, & Mohrman, 1989). One of the most developed and referenced approaches to employee involve- ment is Edward E. Lawler’s (1986, 1992, 1996) “high-involvement work processes” (HIWP). It derives from extensive quality of work–life studies of innovative work systems in the 1970s and was further refined in field and survey studies over the next decades. Lawler’s model defines employee involvement in terms of four mutually reinforcing attributes—power, information, rewards, and knowledge. Power refers to the use of various practices such as participative decision making and job enrich- ment that provide employees a degree of control in decisions that affect their work. Information involves updating employees about company and organization-unit goals as well as providing timely performance feedback. Rewards have to do with the use of contingent reward systems that link compensation, promotions, and rec- ognition to individual, group, and organization performance. Knowledge refers to support for employee skill development through means such as training, job rota- tion, and supervisory coaching. Lawler proposed that the four attributes operate syn- ergistically, such that high involvement can only be attained by enhancing all of them together. Thus, high involvement exists to the extent that employees have the power to make work-related decisions, the requisite information and knowledge to make good decisions, and the rewards for increased performance resulting from good decision making. To achieve high involvement at the organization level, these attributes need to be applied to employees at all levels of the hierarchy. HIWP is proposed to increase organization performance through its positive effects on employee motivation and expertise.

Lawler’s approach to employee involvement has received extensive conceptual and practical attention in the academic and practitioner literature. Empirical assess- ment of the HIWP model has been less extensive, however. Research on the HIWP model has focused mainly on performance outcomes, generally showing that higher levels of HIWP are related in the expected direction to measures of organizational effectiveness, including return on investment, turnover, job performance, and work stress (Butts, Vandenberg, DeJoy, Schaffer, & Wilson, 2009; Lawler et al., 1989; Mackie, Holahan, & Gottlieb, 2001; Macky & Boxall, 2008; O’Neill, Feldman, Vandenberg, Dejoy, & Wilson, 2011; Riordan, Vandenberg, & Richardson, 2005; Vandenberg, Richardson, & Eastman, 1999). Considerably less research has assessed two central features of Lawler’s model: the mechanisms through which HIWP achieves performance results and the synergistic nature of the four HIWP attributes. Only two studies have examined processes mediating the relationship between HIWP and organizational performance, and both focused on attitudinal factors such as employee morale and psychological empowerment (Butts et al., 2009; Vandenberg

at University of Maryland Baltimore Health Sci & Hum Serv Lib on September 27, 2016jab.sagepub.comDownloaded from

Kizilos et al. 415

et al., 1999). Initial evidence shows that these attitudes play a mediating role in how HIWP influences performance, suggesting that motivation accounts for some of the performance effects. Only one study has assessed whether the four HIWP attributes operate as a mutually reinforcing set, and preliminary results support the synergistic relation among the attributes (Vandenberg et al., 1999). Further empirical assess- ment of these two key elements of Lawler’s model would add significantly to knowl- edge of how HIWP works and to the scientific credibility of this approach to employee involvement.

Our study addresses both these issues. We examine a behavioral mechanism through which HIWP is hypothesized to achieve performance effects, organizational citizenship behavior (OCB). OCB involves “performance that supports the social and psychological environment in which task performance takes place” (Organ, 1997, p. 95). In contrast to formally prescribed job behavior, OCB is more discretionary and goes beyond what is organizationally required and enforced. It involves cooperative, shar- ing, and helping behaviors that are intended to benefit the organization. OCB is directed at an organizationally relevant constituency, such as coworkers, suppliers, and customers (Borman & Motowidlo, 1993; Brief & Motowidlo, 1986; Organ, 1988, 1997; Organ, Podsakoff, & MacKenzie, 2006; Van Dyne, Cummings, & Parks, 1995). Although OCB has not been studied in the context of the HIWP model, research on OBC’s antecedents and consequences suggests that it may play a key role in how HIWP affects organization performance (Arthur, 1994; Gong, Chang, & Cheung, 2010; Organ et al., 2006; Podsakoff & MacKenzie, 1994; Snape & Redman, 2010; Sun, Aryee, & Law, 2007).

Our focus on OCB is intended to complement existing attitudinal studies of HIWP mediating effects in two ways. First, HIWP encourages employees to engage in discre- tionary behaviors that benefit the organization. Because these problem-solving, inno- vating, and helping behaviors can promote high performance, they are a potentially important explanation for how HIWP increases organization performance. Second, in contrast to employee attitudes, work behaviors are more directly observable and proxi- mate determinates of organization performance, and thus can potentially offer a more direct account of HIWP’s performance effects than attitudes.

Our study also examines the second main feature of Lawler’s model, whether the four HIWP attributes operate as a distinct bundle of mutually reinforcing elements. This involves methodological considerations of how the four attributes can be mea- sured to account for interrelationships among them and their combined synergistic quality. We operationalize the synergistic nature of the attributes as a higher order construct, which is meant to capture the meaning of HWIP through the common forces underlying the four attributes (Vandenberg et al., 1999). We assess whether this method is statistically valid and produces results that are consistent with Lawler’s conception of how the HIWP attributes work together to create opportunities for employee involvement.

Our study conceptualizes and examines key variables and relationships among them at the collective or organization-unit level, where the performance effects of

at University of Maryland Baltimore Health Sci & Hum Serv Lib on September 27, 2016jab.sagepub.comDownloaded from

416 The Journal of Applied Behavioral Science 49(4)

HIWP and OCB are most likely to occur. We measure the HIWP attributes, OCB, and organization performance using data from 143 business units in a large consumer- products company. The firm implemented a comprehensive employee-involvement program approximately 2 years prior to our study. The primary reason for the pro- gram was to improve customer service, which is essential to competitive success in the consumer-products industry. The data were part of the company’s ongoing assess- ment of this program. The business units are responsible for the sales and distribution of the firm’s products in a geographical area. They are relatively stand-alone busi- nesses with substantial control over their performance outcomes; therefore, they are a good setting to examine OCB’s role in mediating how HIWP affects performance. This setting also affords the opportunity to examine HIWP in an industry in which employee involvement contributes to a key source of competitive advantage.

We begin with a theoretical explanation of the proposed linkages between HIWP and OCB and between OCB and organization-unit performance, and derive relevant hypotheses. Next, we explain the methods used in our study and present the results. Finally, we discuss the findings and draw implications for further research.

Relationship Between HIWP and OCB

Although the linkage between HIWP and OCB has not been studied directly, there is indirect evidence to support this relationship. Various organizational practices that are consistent with HIWP attributes have been shown to be positively related to OCB. These include leadership that is participative, empowering, and rewarding of perfor- mance (MacKenzie, Podsakoff, & Rich, 2001; Schnake, Cochran, & Dumler, 1995); favorable organizational climate that promotes participation and employee support (Niehoff & Moorman, 1992; Schneider, 1987); perceived organizational support for task performance and contingent rewards (Rhoades & Eisenberger, 2002); and enriched work that provides high levels of employee autonomy and performance feedback (Van Dyne, Graham, & Dienesch, 1994). Various explanations have been offered for how these practices promote OCB. Participative and supportive leaders can serve as role models for citizenship behavior and result in followers reciprocating OCB (George & Bettenhausen, 1990; Smith, Organ, & Near, 1983). Participative leadership may lead employees to characterize the organization as supportive (Shore & Wayne, 1993), covenantal (Van Dyne et al., 1994), cooperative (Puffer, 1987) and fair (Niehoff & Moorman, 1992; Schneider, 1987). Consequently, members may be more willing to exert effort on behalf of the organization without expecting direct personal benefits. Enriched forms of work may lead to increased motivation and to feelings of personal responsibility for work outcomes (Pearce & Gregersen, 1991). As a result, employees may be willing to go beyond formally specified duties to accomplish work goals (Becker, 1992; Hatcher, Ross, & Collins, 1989; O’Reilly & Chatman, 1986; Pearce & Gregersen, 1991; Snape & Redman, 2010). Enriched work environments tend to place weaker constraints on employee behaviors and thus can create more opportunities to engage in discretionary behaviors, such as OCB

at University of Maryland Baltimore Health Sci & Hum Serv Lib on September 27, 2016jab.sagepub.comDownloaded from

Kizilos et al. 417

(Schneider, 1983). Contingent rewards can lead employees to engage in OCB to achieve higher levels of performance and the rewards that follow (Organ et al., 2006).

This stream of research offers empirical support for the relationship between HIWP and OCB and suggests plausible reasons for how specific HIWP attributes can lead to those behaviors. However, the majority of this research has focused on single organi- zational practices rather than bundles of interrelated practices. Also, the practices and their relationship to OCB have been studied mainly at the individual level rather than at the organization or unit level.

Concepts related to norm development and to social exchange can help explain how the HIWP attributes work together to promote OCB at the organization-unit level (Ehrhart & Naumann, 2004; Gong et al., 2010; Snape & Redman, 2010; Sun et al., 2007; Takeuchi, Lepak, Wang, & Takeuchi, 2007). HIWP comprises a distinct bundle of attributes that signals to employees what is important and valued in the organization and what they can expect from the employment relationship (Bowen & Ostroff, 2004). It sends clear and consistent messages to employees that their participation in decision making is important to organization performance, they will be supported and listened to, resources will be expended to enhance their skills and expertise, they will receive relevant and timely information, and their rewards will be fair and tied closely to per- formance (Lawler, 1986). These messages comprise what Bowen and Ostroff (2004) have called a “strong human resource management system” that is likely to result in shared perceptions among employees about the organization and its practices. Those shared perceptions, in turn, can inform the development of social norms that infor- mally specify and control what work behaviors are acceptable and unacceptable (Ehrhart & Naumann, 2004). Whether norms promote behavior aimed at benefiting the organization depends on the nature of the exchange relationship between the orga- nization and its members (Gong et al., 2010; Sun et al., 2007).

When employees experience that they are in a high-quality exchange relationship in which the organization provides them with sufficient valued resources (i.e., inducements), they are likely to reciprocate by developing norms that promote posi- tive organizational behaviors (i.e., contributions; Blau, 1964; March & Simon, 1958). HIWP fosters an exchange relationship characterized by mutuality and sup- port between the organization and its members. It provides employees with both monetary and social rewards, such as recognition, work challenge, skill enhance- ment, and social support. Employees are likely to place a high value on these mul- tiple rewards, and consequently, feel obligated to reciprocate by engaging in discretionary work behaviors that benefit the organization, such as OCB. Thus, norms supporting OCB provide employees with a powerful social mechanism for repaying their obligations to the organization. They help employees sustain an exchange relationship with the organization in which the needs of both the organiza- tion and the employees are jointly satisfied.

Hypothesis1: HIWP will be positively related to OCB at the organization-unit level.

at University of Maryland Baltimore Health Sci & Hum Serv Lib on September 27, 2016jab.sagepub.comDownloaded from

418 The Journal of Applied Behavioral Science 49(4)

Relationship Between OCB and Organization-Unit Performance

OCB has attracted considerable attention largely because of its assumed relationship to organization performance (cf. Organ et al., 2006). Although early research found weak support for this relationship at the individual level (George, 1990; Puffer, 1987), a grow- ing number of studies that aggregate OCB to the organization or unit level provide fairly consistent evidence for OCB’s positive performance effects. Studies have found positive relationships between various measures of OCB and multiple indicators of organization performance, including sales, customer satisfaction, quantity and quality of produc- tion, financial efficiency, turnover, and profits (cf. Dunlop & Lee, 2004; George & Bettenhausen, 1990; Koh, Steers, & Terborg, 1995; Koys, 2001; Podsakoff & MacKenzie, 1994; Sun et al., 2007). These findings were obtained in a variety of organizational set- tings (e.g., restaurants, paper mills, retail stores, insurance agencies, hotels, pharmaceuti- cal sales teams, and secondary schools) suggesting their general applicability across manufacturing and service organizations.

So far, no commonly accepted explanation for these positive performance effects has emerged. A variety of plausible reasons have been proposed for how OCB influ- ences organization performance (cf. Organ et al., 2006; Podsakoff & MacKenzie, 1997). One explanation focuses on an organization’s social capital. It argues that OCB can increase social capital by promoting more helpful, smooth, and efficient interac- tions among organization members. This can reduce coordination and control costs (e.g., time and energy spent addressing conflicts, monitoring behavior, and managing task interdependency) and free up valuable resources for more productive purposes (cf. Borman & Motowidlo, 1993; Dovidio, Piliavin, Schroeder, & Penner, 2006). Another reason addresses an organization’s human capital. It suggests that OCB can enhance human capital by members sharing their skills, expertise, and tacit knowledge with each other. This can lead to a more talented, flexible workforce that is likely to be more productive (cf. Evans & Davis, 2005; Kim & Gong, 2009). Still another explana- tion involves employee motivation. It proposes that engaging in OCB can be intrinsi- cally motivating because it affords employees opportunities to experience more freedom, variety, and feedback in their work behaviors. This can lead them to exert more effort to perform and to help others to do the same (cf. Finkelstein, 2011; Snape & Redman, 2010). When taken together, these explanations suggest that OCB increases an organization’s stockpile of social and human capital and motivates its members, which in turn, contribute to organization performance.

Hypothesis 2: OCB will be positively related to organization-unit performance.

Method

Description of Research Setting

Our data were collected from a large global corporation headquartered in the United States that manufactures, markets, and distributes a limited range of perishable (i.e.,

at University of Maryland Baltimore Health Sci & Hum Serv Lib on September 27, 2016jab.sagepub.comDownloaded from

Kizilos et al. 419

short shelf life) consumer products. The company faces heavy competition in an indus- try with essentially substitutable products that customers can discontinue buying with little adverse impact. Because profitability is tied closely to the volume of product sold, special emphasis is placed on providing excellent customer service to sustain and build the customer base. About 2 years prior to our study, a company-wide employee- involvement program was implemented to improve customer service. It included giv- ing frontline service providers more autonomy and control over the total servicing of customer accounts; training them in customer-service skills and process-improvement techniques; symbolically inverting the managerial hierarchy by putting customers at the top, frontline service personnel next, and management at the bottom; changing manage- rial roles from directing and controlling to supporting and coaching with the attendant training and support that is required to implement this change; and scheduling regular information-sharing sessions to ensure that all employees keep up-to-date on company performance, industry dynamics, and changes in the marketplace. Because the firm’s existing reward practices, which included employee stock ownership, individual-based merit pay, and individual and team recognition, were already consistent with employee involvement practices, no significant changes were made to them.

Our study was sanctioned by the company as part of a larger effort to assess the long-term effectiveness of the employee involvement intervention. All data were gath- ered from domestic operations, which accounted for about half of the firm’s revenues of almost $66 billion in 2011. Domestic operations are divided into 15 geographically dispersed business units. These business units are further divided into 115 market units, focused on local market areas. Each market unit is further divided into a number of locations, each responsible for servicing all the commercial accounts within a defined set of territories. Organizations at the location level generally include all the component functions necessary for the production, distribution, and sales of the com- pany’s products. These different functions serve common customers and thus must coordinate their work to deliver excellent customer service. Because the location level is where the performance effects of HIWP are expected to be most evident, it was used as the focal echelon (Rousseau, 1985) of our study and hereafter referred to as “orga- nization unit.”

Focusing on organization units within a single corporation provides certain advan- tages. Measures of HIWP and OCB are more likely to have common meaning across organization units in a single organization than across units in multiple organizations. The organization units in our study used the same performance metric, thus creating a unique opportunity to assess performance in a way that is meaningful to the members of each organization unit and common to a large sample of units. Focusing on perfor- mance variation across units within a single company also controls for a number of alternate explanations for organization performance, such as strategy, technology, and market conditions. The organization units in our study performed similar work in simi- lar environments with similar strategies.

A potential problem of studying units within a single organization is restriction in the range of the variables measured. Because all organization units within the com- pany were involved in the same employee involvement program, members of different

at University of Maryland Baltimore Health Sci & Hum Serv Lib on September 27, 2016jab.sagepub.comDownloaded from

420 The Journal of Applied Behavioral Science 49(4)

units could be expected to have similar perceptions of HIWP and OCB, thus providing relatively little variance in these measures across the units. Conversations with senior executives and human resource administrators suggested that this was unlikely to pose a significant problem in our study. All organization units received the same general guidelines, training, and support for implementing the various features of the employee involvement program. Yet because the units were geographically dispersed and tradi- tionally operated with a good deal of autonomy, they were given considerable leeway in how quickly and extensively they implemented the employee involvement prac- tices. Thus, there was variation across the units in the extent to which employee involvement was implemented.

Complete data were available for 234 of 248 organization units geographically located throughout the United States. They ranged in size from 3 to 744 employees, with a mean of 129 and a median of 65. The majority (52%) of the employees in the organiza- tion units performed sales and distribution work, with the remaining members involved in production (25%), customer service (10%), and various administrative tasks (13%).

Measures

HIWP and OCB. These measures were obtained as a convenience sample from the company’s more extensive survey data collected from employees as part of an ongo- ing human resource assessment process. The firm’s human resource staff administered the surveys on-site at each organization unit during a 2-week period, and employees were guaranteed confidentiality. Seventy-eight percent of the total population of full- time employees participated in the survey. Of the more than 12,000 full-time employ- ees responding, 9,953 employees in 234 organization units provided complete data and were retained for further analyses.

The survey consisted of 150 items, of which 21 were relevant to our study, all with responses measured on a 5-point Likert-type agree/disagree format. The items and their respective scales appear in the appendix to this article.

HIWP was measured with 12 items, with 3 of them for each of the four HIWP attri- butes. The items have good face validity and are similar to measures used in other stud- ies of HIWP (Lawler, Mohrman, & Ledford, 1992; Mackie et al., 2001; Vandenberg et al, 1999). Cronbach’s alpha reliabilities for the HIWP scales were acceptable with the exception of the information scale, which was marginal (Power = .79; Rewards = .84; Knowledge = .83; Information = .63).

OCB was operationalized with relevance to the situation studied. In the organiza- tion units of our study, two major constituencies were pertinent targets for OCB: team members and customers. OCB directed at team members supported team functioning and interaction among members performing separate yet related tasks to sell, produce, and distribute products to a common set of customers. These tasks were relatively standardized and included functional job descriptions. Customer-oriented OCB con- tributed to sustaining a supportive and responsive relationship between the organiza- tion unit and its customers. Nine items were used to measure OCB, four for

at University of Maryland Baltimore Health Sci & Hum Serv Lib on September 27, 2016jab.sagepub.comDownloaded from

Kizilos et al. 421

team-directed OCB and five for customer-directed OCB. The items have good face validity and are similar to measures used in other studies of OCB (see the appendix in Organ et al., 2006, for an excellent review of these different OCB measures).

Although the OCB items derived from a convenience sample and were not created explicitly to measure OCB, they capture the essence of OCB as it was enacted in the organization units. The four items related to team-directed OCB mainly involve dis- cretionary behaviors that exceed what is formally prescribed in the organization’s job descriptions. These cooperative, sharing, and helping behaviors are aimed at improv- ing team functioning and member expertise. The five items measuring customer- directed OCB also involve discretionary behaviors that surpass what the organization formally requires for customer service. The wording of these items shows that these behaviors are intended to “exceed” customers’ expectations, “anticipate” their future needs, and “immediately” resolve their problems, all aspects of customer service that suggest going beyond what the organization can formally specify and require. Finally, both kinds of OCB, team directed and customer directed, are only indirectly tied to organization rewards, such as employee stock ownership, individual merit pay, and individual and team recognition. This is consistent with an earlier, more restrictive conception of OCB as discretionary behavior “not directly or explicitly recognized by the formal reward system” (Organ, 1988, p. 4).

Cronbach’s alpha reliabilities of the two OCB scales were acceptable (Customer- Directed OCB = .85; and Team-Directed OCB = .83). The referent used in these nine items measuring OCB is the “workgroup,” which is consistent with the unit-level con- ception of OCB used in our study (Chan, 1998; Kozlowski & Klein, 2000). Ehrhart and Naumann (2004) argued that when survey data from individual group members uses the group as the referent for OCB, then aggregation of that data to the group level is an accurate description of the group’s OCB norms.

Organization-Unit Performance. Based on personal conversations with a number of the company’s top executives, the most relevant indicator of organization-unit perfor- mance was the actual volume of products sold by the unit, standardized by the unit’s size. This measure reflects the efficiency with which human resources are translated into sales volume (Belcher, 1987; Mitchell, Lewin, & Lawler, 1990). This organization- unit performance measure was taken directly from company records tracking each unit’s yearly sales over 13 four-week accounting periods. Because the survey data for the HIWP and OCB measures were collected during the eighth period, the perfor- mance data for the last five periods in the year were used to calculate organization-unit performance. This time-ordering is consistent with the proposed direction of relation- ships from HIWP to OCB to organization-unit performance.

Aggregating HIWP and OCB Data to the Organization-Unit Level

Because our study explores relationships between HIWP, OCB, and performance at the organization-unit level, individuals’ ratings of HIWP and OCB were aggregated to

at University of Maryland Baltimore Health Sci & Hum Serv Lib on September 27, 2016jab.sagepub.comDownloaded from

422 The Journal of Applied Behavioral Science 49(4)

the organization-unit level. Such aggregation is justified when the data indicate sub- stantial within-setting agreement on the assessed characteristic (Schneider & Bowen, 1985). Agreement suggests that individuals have a shared perception of the character- istic that can be attributed to their membership in a common organization or unit (James, 1982; James, Demaree, & Wolf, 1993). We used James et al.’s (1993) measure of within-setting agreement to calculate the level of organization-unit perceptual agreement (rwg(j)) on the six survey scales. A common yet arbitrary rule of thumb is rwg(j) values of .70 or greater provide empirical justification for aggregation (Lance, Butts, & Michels, 2006). We used a more lenient .65 value for aggregating survey data to the organization-unit level because member responses within an organization unit needed to show agreement at or above that level on all six survey measures, which is a relatively stringent cutoff requirement. The obtained rwg(j) values on the six scales ranged from 0 to 1.0. A total of 143 of the 234 organization units showed agreement at or above the .65 level on all six of the survey measures, thus justifying an organiza- tion-unit level interpretation of their aggregate scores. These 143 organization units were retained and used to test our hypotheses.

Testing the HIWP–OCB–Organization-Unit Performance Relationship

We analyzed the relationships between HIWP and OCB and between OCB and organization-unit performance simultaneously as well as the direct connection between HIWP and performance, which is essential for assessing mediating effects. The covari- ance matrix appearing in Table 1 was analyzed using a LISREL combined structural equation model that simultaneously estimates both measurement error and structural path components and adjusts the value of structural coefficients for measurement error

Table 1. Descriptive Statistics, Correlations, and Covariances.

Mean SD 1 2 3 4 5 6 7

1. Power 3.72 0.28 (.76) .07 .04 .02 .04 .03 .17 2. Rewards 3.60 0.31 .78a (.09) .05 .02 .04 .03 .23 3. Knowledge 3.55 0.25 .60 .66 (.06) .02 .02 .03 .21 4. Information 3.97 0.18 .42 .50 .52 (.03) .02 .02 .13 5. Team-directed

OCB 3.69 0.23 .64 .61 .47 .42 (.05) .03 .18

6. Customer-directed OCB

3.72 0.20 .55 .58 .54 .54 .62 (.04) .21

7. Organization-unit performance

12.52 3.55 .17 .22 .24 .20 .22 .31 (12.58)

Note. OCB = organizational citizenship behavior. Entries above the diagonal are sample covariances, the diagonal elements are variances, and entries below the diagonal are correlations. All values are based on N = 143. a. Correlations greater than .17 are significant at .05; correlations greater than .21 are significant at .01; correlations greater than .27 are significant at .001.

at University of Maryland Baltimore Health Sci & Hum Serv Lib on September 27, 2016jab.sagepub.comDownloaded from

Kizilos et al. 423

(Jöreskog & Sörbom, 1999). The model was estimated using the following matrices: Λy, Γ, Φ, Ψ, and Θε. HIWP and OCB were modeled as two second-order latent factors: ξ1 and ξ2. HIWP was indicated by four first-order factors (η1, η2, η3, and η4): power, information, rewards, and knowledge. OCB was indicated by two first-order factors (η5 and η6): team-directed OCB and customer-directed OCB. Twelve items were used as indicators of the four HIWP attributes, nine items were used as indicators of the two OCB factors, and a single item was used to indicate organization-unit performance (η7). Altogether, these specifications resulted in a 22 × 7 Λy matrix, with specific ele- ments set free to allow each indicator to load only on a single predetermined factor. One indicator on each of the six HIWP and OCB measures was assigned a value of 1 to serve as a reference variable (Jöreskog & Sörbom, 1999). Because performance was assessed through a single archival measure, the indicator associated with this item was set to 1.0 and the corresponding element of the error matrix (θε) was set to 0, indicating that no measurement error was estimated for this indicator. The major model specifica- tions are summarized in graphic form in Figure 1.

Assessing the fit of a structural equations model to observed data involves compari- son of the relationships implied by the hypothesized covariance matrix, Σ, with the relationships observed in the sample covariance matrix, S. This comparison assesses the likelihood that S would be observed if the true relationships producing the data were as specified in Σ. Large differences between the observed and theoretically speci- fied matrices suggest that the specified model is not a good fit to the data (i.e., the theoretical model is not supported by the data). There are numerous indexes for assess- ing model fit, each having certain strengths and weaknesses (cf. Mulaik et al., 1989). We used three commonly accepted indexes: the comparative fit index (CFI; Bentler, 1990; Gerbing, Hamilton, & Freeman, 1994), the incremental fit index (IFI; Bollen, 1989) and the nonnormed fit index (NNFI; Bentler & Bonnett, 1980). Values of these fit indexes generally range from 0 to 1, with values close to 1.0 indicating a good fit. It is generally suggested that models obtaining values of .90 on these indexes represent a good fit (Bentler, 1990). These indexes, along with the χ2 and the χ2/df ratio (Wheaton, Muthan, Alwin, & Summers, 1977), were used to assess the overall fit of the model to the data.

Results

The overall fit of the model to the data was near the acceptable levels on each of the fit measures used in this study: CFI = .86; IFI = .86; NNFI = .84. Although the χ2 = 563.31, df = 202, p < .001 was large, adjustment for the degrees of freedom yielded an acceptable value (χ2/df = 2.79; Fink & Monge, 1985). Overall, the model explained 11% of the variance in organization-unit performance. Although the initial model attained a reasonable overall fit, a series of exploratory respecifications was performed to develop the best-fitting model (Bentler & Bonnett, 1980; Fink & Monge, 1985). When new measures of constructs are used, it is generally acknowledged that the mea- surement portion of an initial model may require respecification in order to obtain an acceptable fit to the data (Anderson & Gerbing, 1988).

at University of Maryland Baltimore Health Sci & Hum Serv Lib on September 27, 2016jab.sagepub.comDownloaded from

424

T ea m -

D ir ec te d

O C B

η 5

C u st o m er -

D ir ec te d

O C B

η 6

O rg a n iz a ti o n -

U n it

P er fo rm a n ce

η 7

H IW P

ξ 1

γ 1 1 γ 2 1

γ 3 1

γ 4 1

O C B

ξ 2

y 1 2

y 1 1

y 1 0

y 9y 8y 7y 6y 5y 4y 3y 2y 1 θ ε

1, 1

θε 2, 2

θε 3, 3

θε 4, 4

θ ε 5, 5

θε 6, 6

θε 7, 7

θε 8, 8

θε 9. 9

θε 10 ,1 0

θε 11 ,1 1

θε 12 ,1 2

P o w er

η 1

In fo rm a ti o n

η 2

R ew a rd s

η 3

K n o w le d g e

η 4

γ 5 2

γ 6 2

γ 7 2

φ 2 1

y 1 3

y 1 4

y 1 5

y 1 6

y 1 7

y 1 8

y 1 9

y 2 0

y 2 1

y 2

2

θε 14

,1 4

θε 15

,1 5

θε 16

,1 6

θε 13

,1 3

θε 18

,1 8

θε 19

,1 9

θε 20

,2 0

θε 17

,1 7

θε 21

,2 1

θε 22

,2 2

λy 1

,1

λy 2

,1

λy 3

,1

λy 4

,2

λy 5

,2

λy 6

,2

λy 7

,3

λy 8

,3

λy 9

,3

λy 1

0 ,4

λy 1

1 ,4

λy 1

2 ,4

λy 1

3 ,5

. . .

. . .λ

y 1

6 ,5

λy 1

7 ,6

. . .

. . .λ

y 2

1 ,6

F ig

u re

1 .

M o de

l r el

at in

g H

IW P

to o

rg an

iz at

io n-

un it p

er fo

rm an

ce t

hr o ug

h O

C B

. N

ot e.

H IW

P =

h ig

h- in

vo lv

em en

t w

o rk

p ro

ce ss

es ; O

C B

= o

rg an

iz at

io na

l c it iz

en sh

ip b

eh av

io r.

at University of Maryland Baltimore Health Sci & Hum Serv Lib on September 27, 2016jab.sagepub.comDownloaded from

Kizilos et al. 425

Based on conceptual considerations and statistical feedback, we performed six respecifications of the measurement part of the model as shown in Table 2. The first three respecifications involved freeing an initially fixed pathway between a HIWP survey item and a HIWP factor, so the item could cross-load on more than one HIWP factor. This resulted in one item from the knowledge factor cross-loading on the rewards factor (K1), one item from the information factor cross-loading on the knowl- edge factor (I3), and one item from the rewards factor cross-loading on the knowledge factor (R3). These findings suggest that the survey respondents may have viewed training as leading to rewards, and knowledge acquisition as helpful to understanding information about the firm’s goals and market. The last three respecifications involved eliminating particular OCB survey items from the two OCB factors. This elimination was based on an item’s poor statistical results, such as being associated with more than one large residual covariance. As shown in Table 2, we eliminated one item from the team-directed OCB factor (T2) and two items from the customer-directed OCB factor (C3, C4). The poor statistical performance of these items may have been caused by ambiguous wording of the items or from organizational constraints that made the behaviors described by the items less subject to discretionary control. For example, two of the items involved solving customer problems immediately and anticipating customer future needs. Employees may have felt that responding effectively to these temporal demands was constrained by their own anticipatory abilities, availability of

Table 2. Summary of Results From Exploratory Respecification of the LISREL Model.

Model Cross-load

added Dropped

item CFI IFI NNFI χ2/df χ2 difference,

p χ2, df, p

Null model — — — — — 11.89 — 2734.41, 230, p < .001

Initial model — — .86 .86 .84 2.79 — 563.31, 202, p < .001

Respecification 1 K1a with rewards

— .87 .87 .85 2.65 31.43, p < .001

531.88, 201, p < .001

Respecification 2 I3 with knowledge

— .88 .88 .86 2.52 28.71, p < .001

503.17, 200, p < .001

Respecification 3 R3 with knowledge

— .88 .88 .87 2.47 12.19, p < .01

490.98, 199, p < .001

Respecification 4 — T2 .90 .90 .88 2.29 — 411.95, 180, p < .001

Respecification 5 — C4 .91 .91 .89 2.15 — 343.99, 160, p < .001

Final model — C3 .92 .92 .90 2.18 — 309.60, 142, p < .001

Note. CFI = comparative fit index; IFI = incremental fit index; NNFI = normed fit index; HIWP = high-involvement work processes; OCB = organizational citizenship behavior; HIWP = high-involvement work processes. a. Survey items measuring HIWP and OCB are numbered the same as they appear next to their descriptions in the appendix: Power (P1, P2, P3), Information (I1, I2, I3), Rewards (R1, R2, R3), Knowledge (K1, K2, K3), Team-Directed OCB (T1, T2, T3, T4), and Customer-Directed OCB (C1, C2, C3, C4, C5).

at University of Maryland Baltimore Health Sci & Hum Serv Lib on September 27, 2016jab.sagepub.comDownloaded from

426 The Journal of Applied Behavioral Science 49(4)

relevant information, and flexibility of company resources. These constraints would limit employees’ capacity to engage in OCB on these dimensions of customer service.

In sum, these modifications to the measurement portion of the initial model resulted in a final, best-fitting model that showed improvement on all the fit indexes: CFI = .92, IFI = .92, and NNFI = .90. The χ2 of 309.60, df = 142, though significant (p < .001), translated into an acceptable χ2/df of 309.60/142 = 2.18. The values for all estimated paths in the model were stable across all the model respecifications, and the structural coefficient estimates were virtually identical across all the models. The results of the final model appear in Figure 2 and are described below.

The operationalization of HIWP as a multidimensional construct indicated by four specific dimensions—power, information, rewards, and knowledge—was supported by their four significant gamma coefficients (γ11 = .25, t = 10.10, p < .001; γ21 = .07, t = 3.9, p < .01; γ31 = .23, t = 8.92, p < .001; γ41 = .17, t = 7.50, p < .001). The squared multiple correlation coefficients (SMCs) for the endogenous variables provided fur- ther support for the HIWP measure, with the multiple equations explaining a reason- able amount of the variance in these factors (SMCpower = .82; SMCinformation = .29; SMCrewards = .84; and SMCknowledge = .41). The SMCs for the observed vari- ables indicated that each first-order HIWP factor attained acceptable levels, with val- ues ranging from .31 to .85; all the t values associated with the indicators were significant.

The loadings of the two first-order OCB factors on the second-order OCB measure were quite good (γ52 = .20, t = 8.52, p < .001, and γ62 = .12, t = 7.99, p < .001). A large portion of the variance in the first-order factors was explained by their defining equa- tions, with SMC customer-directed OCB = .70 and SMC team-directed OCB = .69. The SMCs for the observed indicators of the two first-order OCB factors were gener- ally high, with values ranging from .50 to .85; all the t values associated with the indicators were significant.

Inspection of the matrices of error covariances supported the assumptions of gener- ally uncorrelated errors among the indicators (Θε) and among the latent variables (Ψ).

The results support both hypotheses explored in this study. HIWP was positively related to OCB (Hypothesis 1: Φ21 = .87, t = 18.83, p < .001), and OCB was positively related to organization-unit performance (Hypothesis 2: γ72 = 1.04, t = 3.82, p < .01). Moreover, the modification index for the omitted direct path from HIWP to organization- unit performance was .91. According to Jöreskog and Sörbom (1999), values greater than 4 suggest that the inclusion of the path may improve the model. Thus, our find- ings suggest that the relationship between HIWP and organization-unit performance was completely mediated by OCB.

Discussion and Conclusion

Our results show that HIWP was positively and significantly related to organization- unit performance in a sample of 143 consumer-products organization units.

at University of Maryland Baltimore Health Sci & Hum Serv Lib on September 27, 2016jab.sagepub.comDownloaded from

427

T ea m -

D ir ec te d

O C B

η 5

C u st o m er -

D ir ec te d

O C B

η 6

O rg a n iz a ti o n -

U n it

P er fo rm a n ce

η 7

H IW P

ξ 1 O C B

ξ 2

P 1a

P 2

P 3

I3 I1 I2 R 3

R 1

R 2

K 2

K 1

K 3

.0 4

.0 2

.0 2

.0 3

.0 3

.0 1

.0 2

.0 2

.0 3

.0 1

.0 2

.0 1

P o w er

η 1

In fo rm a ti o n

η 2

K n o w le d g e

η 4

y 2 2

1. 0

.9 5

.8 9 1 .0

1. 19 1. 15

1. 0

1. 05 1. 16

1. 0

.4 8

.9 7

.2 5 b (. 9 1 )c

.0 7 (. 5 4 )

.2 3 (. 9 1 ) .1 7 (. 6 4 )

.8 7 (. 8 7 )

.2 0 (. 8 5 )

.1 2 (. 8 2 )

1 .0 4

(. 3 3 )

1 .0

0 .0

.3 9

R ew a rd s

η 3

.4 8

T 2

C 1

C 2

C 5

.0 1

.0 3

.0 2

T 3

T 1

T 4

.0 4

.0 2

.0 1

1 .0

.9 6

.7 9

1 .0

1 .2 7

1 .1 4

.2 7

D ro p p ed

C 4

D ro p p ed

C 3

D ro p p ed

F ig

u re

2 .

LI SR

EL r

es ul

ts f ro

m f in

al r

es pe

ci fic

at io

n o f th

e m

o de

l r el

at in

g H

IW P

to o

rg an

iz at

io n-

un it p

er fo

rm an

ce t

hr o ug

h O

C B

. N

ot e.

H IW

P =

h ig

h- in

vo lv

em en

t w

o rk

p ro

ce ss

es ; O

C B

= o

rg an

iz at

io na

l c it iz

en sh

ip b

eh av

io r.

a. S

ur ve

y it em

s m

ea su

ri ng

H IW

P an

d O

C B

a re

n um

be re

d th

e sa

m e

as t

he y

ap pe

ar n

ex t

to t

he ir

d es

cr ip

ti o ns

in t

he a

pp en

di x:

P o w

er (

P1 , P

2, P

3) , I

nf o rm

at io

n (I

1, I 2,

I 3)

, R ew

ar ds

( R

1, R

2, R

3) , K

no w

le dg

e (K

1, K

2, K

3) , T

ea m

-D ir

ec te

d O

C B

( T

1, T

2, T

3, T

4) , a

nd C

us to

m er

-D ir

ec te

d O

C B

( C

1, C

2, C

3, C

4, C

5) .

b. A

ll co

ef fic

ie nt

s ar

e si

gn ifi

ca nt

a t, a

t le

as t, t

he p

< .0

1 le

ve l.

c. V

al ue

s in

p ar

en th

es es

a re

s ta

nd ar

di ze

d lo

ad in

gs .

at University of Maryland Baltimore Health Sci & Hum Serv Lib on September 27, 2016jab.sagepub.comDownloaded from

428 The Journal of Applied Behavioral Science 49(4)

As hypothesized, this relationship was mediated through OCB, team-directed and customer-directed. Taken together, the relationships in the model explained about 11% of the variance in organization-unit performance, a modest yet significant amount of explanatory value. To illustrate the magnitude of the relationship between HIWP and organization-unit performance in concrete terms, we split the sample of organization units into thirds based on their values on the HIWP measures and compared the top and bottom thirds on performance. The top third attained a level of performance that was 18.8% higher than that of the bottom third, suggesting that HIWP can have sub- stantial performance consequences for organizations.

Our findings contribute to knowledge about the performance effects of Lawler’s (1986, 1992, 1996) HIWP model. They complement existing attitudinal studies of HIWP’s performance effects by showing that discretionary work behaviors such as OCB also play a significant mediating role in the relationship between HIWP and organization performance. The results are consistent with explanations of OCB’s mediating effects grounded in concepts about norm development and social exchange (Ehrhart & Naumann, 2004; Gong et al., 2010; Sun et al., 2007; Takeuchi et al., 2007). We proposed that HIWP promotes OCB when the four attributes—power, informa- tion, rewards, and knowledge—work together synergistically to create a work envi- ronment where employee knowledge and skills are valued, productively used, and rewarded. Common exposure to this situation, and the strong messages it sends, leads to shared perceptions among employees that they are in a high-quality exchange rela- tionship with the organization. To reciprocate for the economic and social rewards derived from HIWP, employees develop group norms that support OCB to benefit the organization. OCB enhances the organization’s social and human capital and moti- vates its members, which in turn, promote high performance.

We operationalized OCB at the organization-unit level, which is consistent with the view that OCB is the normative mode of behavior in the units studied. Our measures of OCB were directly relevant to achieving the organization units’ strategic goal of superior customer service, which is essential for competitive advantage in the consumer-products industry. Team-directed OCB involved members’ willingness to engage in decision making, helping behavior, and knowledge sharing. Customer- directed OCB included helping customers solve problems, adapting to their changing needs, and devising ideas to exceed their expectations. Excellent customer service was essential to attaining high sales volume, our measure of organization-unit perfor- mance. These measures of OCB and organization-unit performance are responsive to increasing calls to anchor research on the performance effects of employee involve- ment to industry competitive dynamics. Research findings are likely to be more per- suasive and practically relevant when performance measures are directly relevant to the competitive conditions of the organization or unit studied (Boxall & Macky, 2009).

In addition, our results add support to Lawler’s (1986, 1992, 1996) notion that the four HIWP attributes—power, rewards, knowledge and information—are mutually reinforcing and work together to create opportunities for employee involvement. Consistent with preliminary research on the attributes (Vandenberg et al., 1999), our

at University of Maryland Baltimore Health Sci & Hum Serv Lib on September 27, 2016jab.sagepub.comDownloaded from

Kizilos et al. 429

findings confirm that the synergistic nature of the attributes can be captured by opera- tionalizing them as a higher order construct. This treats the attributes as a collective set of mutually reinforcing elements rather than as individual attributes working sepa- rately from each other.

The findings contribute to the generalizability of the HIWP model. They show that the proposed relationships between HIWP, OCB, and organization-unit performance operate in a consumer-products setting where customer service is key to organization success. This complements other studies of HIWP in settings related to life insurance (Riordan et al., 2005; Vandenberg et al., 1999), residential health care (Mackie et al., 2001), and retail sales (Butts et al., 2009; O’Neill et al., 2011). Because we examined the proposed relationships at the organization-unit level, the findings reinforce the value of aggregating data and analyzing relationships at the organizational level where they are most likely to be operative and relevant.

Our findings should be interpreted in the context of the limitations of the study. First, the data are nonexperimental and essentially cross-sectional. Although the sur- vey data were collected at one point in time and the performance data at a later point in time, it is not possible to draw causal inferences from the data in this study. The results are consistent with the theoretical predictions relating HIWP to organization- unit performance through OCB; however, the existence of alternative equivalent mod- els and the observational nature of the data suggest caution in interpreting the results in causal terms. It is possible, for example, that the causal process actually operates in the reverse direction, from organization-unit performance to OCB to HIWP, or that other variables not measured in this study would have yielded the same or better results. It also can be argued that the reported relationships between variables mea- sured across the two time periods in the study were due to a high serial correlation of performance (r = .87, p < .001) rather than a causal relationship. To test for this alter- native explanation, we performed a three-step hierarchical regression analysis to con- trol for past performance in the prediction of future performance. Past performance was entered as the sole predictor of future performance in the initial regression and the obtained R2 was 75.20%. The addition of HIWP resulted in a marginally significant increment to the R2 (R2 = 75.78%, F = 3.28, df = 1,137, p < .10) and the further addi- tion of OCB resulted in an additional marginally significant improvement (R2 = 76.39%, F = 3.51, df = 1,136, p < .10). Although these results are consistent with the causal direction of the proposed relationships in our study, they do not permit causal inference.

Second, despite all the organization units being in the same company, it is possible that differences in their market environments may have influenced the results. The organization unit’s share of the local market, the growth of the local market, and the prices of its products are a few of the factors that may have varied across the organiza- tion units and influenced the results. To examine the possibility that these factors cre- ated an omitted variable bias in our results, we performed a post hoc LISREL analysis with a modified performance variable. We first regressed performance on market share, industry growth, and average price. Next, we used the residuals

at University of Maryland Baltimore Health Sci & Hum Serv Lib on September 27, 2016jab.sagepub.comDownloaded from

430 The Journal of Applied Behavioral Science 49(4)

from this analysis as a modified performance variable in a LISREL analysis. The only differences in these results from those reported in the “Results” section were a slightly reduced overall fit of the proposed model (CFI = .86) and a slight reduction in the SMC for performance (SMCperformance = .10), which was expected because the residualizing procedure reduced the variance in the performance measure.

Third, our results may have limited generalizability to other kinds of OCB and to other types of organizational settings. The specific measures of OCB in this study were selected because they were maximally relevant to the setting (Organ, 1988). Team-directed OCB and customer-directed OCB were aimed at delivering superior customer service. These OCBs may not be relevant to performance in other settings and may not even be considered OCB in others. The particular types of OCB that are relevant to performance can be expected to vary with corporate strategy, the nature of work, industry characteristics, and a host of other contextual features (cf. Datta, Guthrie, & Wright, 2005; Gong et al., 2010). The organizational setting for our study was a large global corporation in the perishable consumer-products industry. The com- pany implemented HIWP to achieve more flexible, rapid, and innovative responses to customer needs, which are essential for competitive advantage in this industry. Generalizing our results to a different organization context should be done with caution.

Fourth, as mentioned in the “Method” section, a potential drawback in studying the performance effects of HIWP in a single organization can be restriction in the range of the variables measured, especially the four HIWP attributes and the two OCB mea- sures. This could result in weak support for hypothesized relationships. The standard deviations for our measures shown in Table 1 are quite modest, ranging between .18 and .31 for the HIWP and OCB measures. This suggests that our findings, which are relatively strong, might be even more robust with data collected from multiple organizations.

Organizations are increasingly using employee-involvement approaches to increase performance and to gain competitive advantage. Our study focused on an important and widely referenced involvement system, Lawler’s HIWP model. Research has found consistent evidence for a positive relationship between HIWP and organization performance. What is needed, however, is research-based understanding of how the four attributes of the HIWP model interrelate and mutually reinforce each other, and in turn, through what mechanisms the attributes contribute to organization performance. Our findings showed that the HIWP attributes function together at a higher order syn- ergistic level that is related to organization-unit performance through employees’ OCB. The results add significant empirical knowledge of how the HIWP model works and achieves performance effects. They reinforce the credibility of Lawler’s model as an evidence-based approach to employee involvement. Future research can examine the variables and relationships of our study longitudinally to permit strong causal inference of the results.

at University of Maryland Baltimore Health Sci & Hum Serv Lib on September 27, 2016jab.sagepub.comDownloaded from

Kizilos et al. 431

Appendix

Survey Items Measuring High-Involvement Work Processes (HIWP) and Organizational Citizenship Behavior (OCB)a

HIWP Power

P1 My manager asks for employees’ input before making a decision that affects them P2 My manager encourages us to take charge of the situation and make decisions P3 My manager helps me develop my own solutions rather than always giving me

the answers

Information

I1 I understand (the Company’s) national business goals and objectives I2 I understand the goals and objectives of my workgroup I3 I receive the information I need to do my job

Rewards

R1 My manager recognizes good performance R2 When my team performs well, all of the team members are recognized R3 My manager rewards people when they improve

Knowledge

K1 I am given an opportunity to improve my skills K2 I receive formal training when I need it K3 I receive the training I need to do my job well

Collective OCB

Team-Directed OCB. “People in my workgroup . . .”

T1 Make sure that other team members are involved in decisions that may affect them T2 Support and act on the group’s decisions T3 Pitch in to help out when necessary T4 Share knowledge or experience that can benefit others in the workgroup

Customer-Directed OCB. “People in my workgroup . . .”

C1 Come up with good ideas to exceed what the customer says he or she wants C2 Act on their ideas to exceed customer expectations

at University of Maryland Baltimore Health Sci & Hum Serv Lib on September 27, 2016jab.sagepub.comDownloaded from

432 The Journal of Applied Behavioral Science 49(4)

C3 Resolve customer problems immediately C4 Anticipate customers’ future needs C5 Are willing to adapt to meet the changing needs of our customers

a. For ease of interpreting the LISRL results, the item numbers are the same as those shown in Table 2 and Figure 2.

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

References

Anderson, J., & Gerbing, D. (1988). Structural equation modeling in practice: A review and recommended two-step approach. Psychological Bulletin, 103, 411-423.

Arthur, J. (1994). Effects of human resource systems on manufacturing performance and turn- over. Academy of Management Journal, 37, 670-687.

Becker, T. (1992). Foci and bases of commitment: Are they distinctions worth making? Academy of Management Journal, 35, 232-244.

Belcher, J. (1987). Productivity plus: How today’s best run companies are gaining the competi- tive edge. Houston, TX: Gulf.

Bentler, P. (1990). Comparative fit indexes in structural models. Psychological Bulletin, 107, 2, 238-246.

Bentler, P., & Bonnett, D. (1980). Significance tests and goodness of fit in the analysis of cova- riance structures. Psychological Bulletin, 88, 588-606.

Blau, P. (1964). Exchange and power in social life. New York, NY: Wiley. Bollen, K. (1989). A new incremental fit index for general structural equation models.

Sociological Methods & Research, 17, 303-316. Borman, W., & Motowidlo, S. (1993). Expanding the criterion domain to include elements of

contextual performance. In N. Schmitt, W. Borman, & Associates (Eds.), Personnel selec- tion in organizations (pp. 71-98). San Francisco, CA: Jossey-Bass.

Bowen, D., & Ostroff, C. (2004). Understanding HRM-firm performance linkages: The role of the “strength” of the HRM system. Academy of Management Review, 29, 203-221.

Boxall, P. and Macky, K. (2009). Research and theory on high performance work systems: Progressing the high-involvement stream. Human Resource Management Journal, 19, 3–23.

Brief, A., & Motowidlo, S. J. (1986). Organizational citizenship behaviors. Academy of Management Review, 11, 710-725.

Butts, M., Vandenberg, R., DeJoy, D., Schaffer, B., & Wilson, M. (2009). Individual reactions to high involvement work processes: Investigating the role of empowerment and perceived organizational support. Journal of Occupational Health Psychology, 14, 122-136.

Chan, D. (1998). Functional relationships among constructs in the same content domain at differ- ent levels of analysis: A typology of composition models. Journal of Applied Psychology, 83, 234-246.

at University of Maryland Baltimore Health Sci & Hum Serv Lib on September 27, 2016jab.sagepub.comDownloaded from

Kizilos et al. 433

Conger, J., & Kanungo, R. (1988). The empowerment process: Integrating theory and practice. Academy of Management Review, 13, 471-482.

Datta, D., Guthrie, J., & Wright, P. (2005). Human resource management and labor productiv- ity: Does industry matter? Academy of Management Journal, 48, 135-145.

Dovidio, J., Piliavin, J., Schroeder, D., & Penner, L. (2006). The social psychology of prosocial behavior. Mahwah, NJ: Lawrence Erlbaum.

Dunlop, P. D., & Lee, K. (2004). Workplace deviance, organizational citizenship behavior, and business unit performance: The bad apples do spoil the whole barrel. Journal of Organizational Behavior, 25, 67-80.

Ehrhart, M., & Naumann, S. (2004). Organizational citizenship behavior in work groups: A group norms approach. Journal of Applied Psychology, 89, 960-974.

Evans, W., & Davis, W. (2005). High-performance work systems and organizational perfor- mance: The mediating role of internal structure. Journal of Management, 31, 758-775.

Fink, E. L., & Monge, P. R. (1985). An exploration of confirmatory factor analysis. In B. Dervin & M. J. Voigt (Eds.), Progress in communication sciences (Vol. 6, pp. 167-197). Norwood, NJ: Ablex.

Finkelstein, M. (2011). Intrinsic/extrinsic motivation and organizational citizenship behavior: A functional approach to OCB. Journal of Psychological Issues in Organizational Culture, 2, 19-34.

Gerbing, D., Hamilton, J., & Freeman, E. (1994). A large-scale second-order structural equation model of the influence of management participation on organizational planning benefits. Journal of Management, 20, 859-886.

George, J. M. (1990). Personality, affect, and behavior in groups. Journal of Applied Psychology, 75, 107-116.

George, J. M., & Bettenhausen, K. (1990). Understanding prosocial behavior, sales perfor- mance, and turnover: A group-level analysis in a service context. Journal of Applied Psychology, 75, 698-709.

Gong, Y., Chang, S., & Cheung, S. (2010). High performance work system and collective OCB: A collective social exchange perspective. Human Resource Management Journal, 20, 119-137.

Hatcher, L., Ross, T., & Collins, D. (1989). Prosocial behavior, job complexity, and suggestion contribution under gainsharing plans. Journal of Applied Behavioral Science, 25, 231-248.

James, L. (1982). Aggregation bias in estimates of perceptual agreement. Journal of Applied Psychology, 67, 219-229.

James, L., Demaree, R., & Wolf, G. (1993). rwg: An assessment of within-group interrater agreement. Journal of Applied Psychology, 78, 306-309.

Jöreskog, K., & Sörbom, D. (1999). LISREL 8: User’s reference guide. Chicago, IL: Scientific Software International.

Kim, H., & Gong, Y. (2009). The roles of tacit knowledge and OCB in the relationship between group-based pay and firm performance. Human Resource Management Journal, 19, 120-139.

Koh, W., Steers, R., & Terborg, J. (1995). The effects of transformational leadership on teacher attitudes and student performance in Singapore. Journal of Organizational Behavior, 16, 319-333.

Koys, D. (2001). The effects of employee satisfaction, organizational citizenship behavior, and turnover on organizational effectiveness: A unit-level, longitudinal study. Personnel Psychology, 54, 101-114.

Kozlowski, S., & Klein, K. (2000). A multilevel approach to theory and research in organiza- tions: Contextual, temporal, and emergent processes. In K. Klein & S. Kozlowski (Eds.),

at University of Maryland Baltimore Health Sci & Hum Serv Lib on September 27, 2016jab.sagepub.comDownloaded from

434 The Journal of Applied Behavioral Science 49(4)

Multilevel theory, research, and methods in organizations: Foundations, extensions, and new directions (pp. 3-90). San Francisco, CA: Jossey-Bass.

Lance, C., Butts, M., & Michels, L. (2006). The sources of four commonly reported cutoff cri- teria: What did they really say? Organization Research Methods, 9, 202-220.

Lawler, E. (1986). High involvement management: Participative strategies for improving orga- nizational performance. San Francisco, CA: Jossey-Bass.

Lawler, E. (1992). The ultimate advantage. San Francisco, CA: Jossey-Bass. Lawler, E. (1996). From the ground up: Six principles for building the new logic corporation.

San Francisco, CA: Jossey-Bass. Lawler, E., & Boudreau, J. (2009). Achieving excellence in human resources management: An

assessment of human resource functions. Stanford, CA: Stanford University Press. Lawler, E., Ledford, G., & Mohrman, S. (1989). Employee empowerment in America. Houston,

TX: American Productivity and Quality Center. Lawler, E., Mohrman, S., & Ledford, G. (1992). Employee involvement and total quality manage-

ment: Practices and results in Fortune 1000 companies. San Francisco, CA: Jossey-Bass. MacKenzie, S., Podsakoff, P., & Rich, G. (2001). Transformational and transactional leadership

and salesperson performance. Journal of the Academy of Marketing Science, 29, 115-134. Mackie, K., Holahan, C., & Gottlieb, N. (2001). Employee involvement management practices,

work stress, and depression in employees of a human services residential care facility. Human Relations, 54, 1065-1092.

Macky, K., & Boxall, P. (2008). High-involvement work processes, work intensification and employee well-being: A study of New Zealand worker experiences. Asia Pacific Journal of Human Resources, 46, 38-55.

March, J., & Simon, H. (1958). Organizations. New York, NY: Wiley. Mitchell, D., Lewin, D., & Lawler, E. (1990). Alternative pay systems, firm performance, and

productivity. In A. Blinder (Ed.), Paying for productivity: A look at the evidence (pp. 15-94). Washington, DC: Brookings Institution.

Mulaik, S., James, L., van Alstie, J., Bennett, N., Lind, S., & Stilwell, C. (1989). Evaluation of goodness-of-fit indices for structural equation models. Psychological Bulletin, 105, 430-455.

Niehoff, B., & Moorman, R. (1992). Justice as a mediator of the relationship between methods of monitoring and organizational citizenship behavior. Academy of Management Journal, 36, 527-556.

O’Neill, O., Feldman, D., Vandenberg, R., Dejoy, D., & Wilson, M. (2011). Values, high- involvement work practices, and business unit performance. Human Resource Management, 50, 541-558.

O’Reilly III, C., & Chatman, J. (1986). Organizational commitment and psychological attach- ment: The effects of compliance, identification, and internalization on prosocial behavior. Journal of Applied Psychology, 71, 492-499.

Organ, D. (1988). Organizational citizenship behavior: The good soldier syndrome. Lexington, MA: Lexington Books.

Organ, D. (1997). Organizational citizenship behavior: It’s construct cleanup time. Human Performance, 10, 85-97.

Organ, D., Podsakoff, P., & MacKenzie, S. (2006). Organizational citizenship behavior: Its nature, antecedents, and consequences. Thousand Oaks, CA: Sage.

Pearce, J., & Gregersen, H. (1991). Task interdependence and extrarole behavior: A test of the mediating effects of felt responsibility. Journal of Applied Psychology, 76, 838-844.

at University of Maryland Baltimore Health Sci & Hum Serv Lib on September 27, 2016jab.sagepub.comDownloaded from

Kizilos et al. 435

Podsakoff, P., & MacKenzie, S. (1994). Organizational citizenship behaviors and sales unit effectiveness. Journal of Marketing Research, 31, 351-363.

Podsakoff, P., & MacKenzie, S. (1997). Impact of organizational citizenship behavior on organi- zational performance: A review and suggestions for future research. Human Performance, 10, 133-151.

Puffer, S. (1987). Prosocial behavior, noncompliant behavior, and work performance among commission salespeople. Journal of Applied Psychology, 72, 615-621.

Rhoades, L., & Eisenberger, R. (2002). Perceived organizational support: A review of the litera- ture. Journal of Applied Psychology, 87, 698-714.

Riordan, C., Vandenberg, R., & Richardson, H. (2005). Employee involvement climate and organizational effectiveness. Human Resource Management, 44, 471-488.

Rousseau, D. (1985). Issues of level in organizational research: Multilevel and cross-level per- spectives. Research in Organizational Behavior, 7, 1-37.

Schnake, M., Cochran, D., & Dumler, M. (1995). Encouraging organizational citizenship: The effects of job satisfaction, perceived equity and leadership. Journal of Management Issues, 7, 209-221.

Schneider, B. (1983). Interactional psychology and organizational behavior. Research in Organizational Behavior, 5, 1-31.

Schneider, B. (1987). The people make the place. Personnel Psychology, 40, 437-453. Schneider, B. & Bowen, D. (1985). Employee and customer perceptions of service in banks:

Replication and extension. Journal of Applied Psychology, 70, 423-433. Shore, L., & Wayne, S. (1993). Commitment and employee behavior: Comparison of affective

commitment and continuance commitment with perceived organizational support. Journal of Applied Psychology, 78, 774-780.

Smith, C., Organ, D., & Near, J. (1983). Organizational citizenship behavior: Its nature and antecedents. Journal of Applied Psychology, 68, 653-663.

Snape, E., & Redman, T. (2010). HRM practices, organizational citizenship behaviour, and performance: A multi-level analysis. Journal of Management Studies, 47, 1219-1247.

Sun, L., Aryee, S., & Law, K. (2007). High-performance human resource practices, citizen- ship behavior, and organizational performance: A relational perspective. Academy of Management Journal, 50, 558-577.

Takeuchi, R., Lepak, D., Wang, H., & Takeuchi, K. (2007). An empirical examination of the mechanisms mediating between high-performance work systems and the performance of Japanese organizations. Journal of Applied Psychology, 92, 1069-1083.

Van Dyne, L., Cummings, L., & Parks, M. (1995). Extra role behaviors: In pursuit of con- struct and definitional clarity (a bridge over muddied waters). Research in Organizational Behavior, 17, 215-285.

Van Dyne, L., Graham, J., & Dienesch, R. (1994). Organizational citizenship behavior: Construct redefinition, measurement and validation. Academy of Management Journal, 37, 765-802.

Vandenberg, R., Richardson, H., & Eastman, L. (1999). The impact of high involvement work processes on organizational effectiveness: A second-order latent variable approach. Group & Organization Management, 24, 300-339.

Wheaton, B., Muthan, B., Alwin, D., & Summers, G. (1977). Assessing reliability and stability in panel models. In D. Heise (Ed.), Sociological methodology (pp. 84-136). San Francisco, CA: Jossey-Bass.

at University of Maryland Baltimore Health Sci & Hum Serv Lib on September 27, 2016jab.sagepub.comDownloaded from

436 The Journal of Applied Behavioral Science 49(4)

Author Biographies

Mark A. Kizilos received his PhD from the Marshall School of Business at the University of Southern California. He is President and Founder of Experience-Based Development Associates, a management development and consulting firm. His extensive research led to the development of the toolkit and book, FrameBreaking Leadership Development.

Chailin Cummings received her PhD from the Marshall School of Business at the University of Southern California. She is currently an assistant professor of strategic management at the College of Business Administration, California State University, Long Beach. Her research interests include organization identify, work engagement, and network analysis of industry for- mation and dynamics.

Thomas G. Cummings is a professor and chair of the department of management and organiza- tion at the Marshall School of Business, University of Southern California. His research inter- ests include designing high-performance organizations, strategic change, and evolution of sci- entific fields. His has authored 23 books and over 70 academic articles, and served as President of the Academy of Management, the leading professional organization for scholars in manage- ment and organization.

at University of Maryland Baltimore Health Sci & Hum Serv Lib on September 27, 2016jab.sagepub.comDownloaded from