SCOTT M. SOLTIS, FILIP AGNEESSENS, ZUZANA ... SOCIAL NETWORK PERSPECTIVE ON TURNOVER INTENTIONS: THE...

A SOCIAL NETWORK PERSPECTIVE

ON TURNOVER INTENTIONS: THE

ROLE OF DISTRIBUTIVE JUSTICE

AND SOCIAL SUPPORT

S C O T T M . S O LT I S , F I L I P A G N E E S S E N S , Z U Z A N A S A S O V O V A , A N D G I U S E P P E ( J O E ) L A B I A N C A

Organizations are increasingly concerned about retaining human talent, par-

ticularly within knowledge-based industries where turnover is expensive.

Our study employs a social network perspective to explore the infl uence of

employees’ formal and informal workplace relationships on their turnover

intentions. We do this in a life sciences organization experiencing employee

turnover at over twice the rate of the industry average. Drawing on extant

work on the effects of distributive justice at work, we argue that employ-

ees who are heavily sought out for advice see themselves as being under-

rewarded for the time and effort that goes into providing advice, thus in-

creasing turnover intentions. Additionally, we argue that employees see the

ability to seek out advice as a form of social support that embeds them into

the organization and decreases intention to quit. By exploring the network

positions of individuals in the workfl ow and advice networks, we demon-

strate that when employees are either providing advice to someone they are

obligated to work with or are able to seek out advice from others who are not

required to work with them, the relationship with turnover intentions is most

intense. We conclude by discussing contributions to the theory and practice

of human resource management.

Keywords: social networks, multiplexity, turnover intentions

Correspondence to: Scott M. Soltis, College of Business Administration, University of Missouri-St. Louis, 246

Express Scripts Hall, One University Blvd., St. Louis, MO 63121-4400, Phone: 314-516-6279, Fax: 314-516-6827,

E-mail: [email protected].

Human Resource Management, July–August 2013, Vol. 52, No. 4. Pp. 561–584

© 2013 Wiley Periodicals, Inc.

Published online in Wiley Online Library (wileyonlinelibrary.com).

DOI:10.1002/hrm.21542

Employee turnover is ubiquitous in or-ganizational life, even in the midst of severe economic recessions where job-less rates are higher than normal. It is among the most frequently studied

topics in the human resource management

literature. Our understanding of what leads to employee turnover and the impacts of that turnover on the organization has increased considerably after decades of research (see Holtom, Mitchell, Lee, & Eberly, 2008; Shaw, 2011, for recent reviews). Turnover often has

562 HUMAN RESOURCE MANAGEMENT, JULY–AUGUST 2013

Human Resource Management DOI: 10.1002/hrm

Informal ties

supplement the

formal structures

of the organization

by providing

other avenues to

overcome problems

that the formal

structure cannot

handle; without

them, organizations

would likely grind to

a halt.

detrimental effects on organizational perfor-mance due to lost human and social capital (Shaw, Duffy, Johnson, & Lockhart, 2005). All told, the employee withdrawal process is a costly one that can consume as much as 17 percent of a company’s pretax income (Sagie, Birati, & Tziner, 2002). These negative organi-zational ramifications have fueled sustained interest in understanding employee turnover. Many models of turnover since Mobley (1977) find that employees do not typically leave an organization on a whim; rather, they fre-quently have turnover intentions that precede actions such as job search, withdrawal behav-iors, or organizational exit, and these turnover intentions are the focus of this study.1

A recent influential theory on turnover, the Job Embeddedness Model (JEM; Mitchell, Holtom, Lee, Sablynski, & Erez, 2001) argues that the “links” an employee has within the work-place can have a strong influence on turnover. Drawing from lit-erature on attachment (March & Simon, 1958), Mitchell and col-leagues (2001, p. 1104) “describe job embeddedness as like a net or a web in which an individual can become stuck.” Our study’s focus is on employees’ social ties (in the JEM referred to as “links”) to other coworkers, whether they are formally prescribed by the organi-zation or emerge informally with-out managerial or organizational inducement. Together these for-mal and informal social ties make up the social web that defines the organization as experienced by the employee every day when coming to work. We contend

that these informal and formal social ties are related to the extent to which an employee harbors turnover intentions, and that some social ties will actually promote feelings of being overwhelmed or under-rewarded, and thereby increase turnover intentions.

Most research examines individuals’ atti-tudes or demographic characteristics as rea-sons for leaving without accounting for the

potential influence of the social structures within which these employees are embed-ded through their social ties (Griffeth, Hom, & Gaertner, 2000). In contrast, we explore how these informal and formal workplace social ties may be related to turnover inten-tions. Exploring informal relationships, such as whom one goes to for work-related help or advice, and formal relationships in the organization, such as whom one is required to work with, is important for theoretical and practical reasons. Formal relationships are important theoretically because employ-ees cannot easily avoid them and may have little agency in determining their existence. Informal workplace ties are crucial to under-standing how work actually gets done in an organization. Informal ties supplement the formal structures of the organization by pro-viding other avenues to overcome problems that the formal structure cannot handle; without them, organizations would likely grind to a halt (e.g., Morey & Luthans, 1991).

Rather than explore these informal and formal relationships separately, as is the case with most network studies, we examine them jointly to determine where overlapping, mul-tiplex ties do and do not occur. Specifically, we argue that the extent to which a social tie is perceived to be voluntary or required by formal workflows can drastically alter the relationship between giving or receiving advice and turnover intentions. For exam-ple, employees giving advice to a coworker even though they do not feel required to do so by the formal workflow design would be interpreted as acts of social support being received by the coworker. Those individuals feeling socially supported might be less apt to want to leave a positive social workplace environment compared to those who are not feeling so supported. Whether or not formal and informal ties are intertwined may also have an effect on the advice giver. Providing advice, particularly to those with whom one feels required to work, takes away from one’s abilities to complete their own tasks and may even reduce the advice giver’s individual per-formance. The frustration that may be associ-ated with feeling as if one is required to give informal advice, especially if one is not being

A SOCIAL NETWORK PERSPECTIVE ON TURNOVER INTENTIONS 563


While the full JEM

includes ties within

the workplace and

those to family,

community, and

other external

entities, research

suggests that those

ties internal to the

workplace have

greater influence

on attitudes such

as satisfaction or

commitment than do

external ties.

rewarded for advice giving, might adversely affect how attached one is to the organization.

Our main contribution is to examine the multiple direct and indirect effects of required and voluntary workplace ties on turnover intentions, which allows us to begin to answer calls for greater attention to the role of social networks in turnover (Dess & Shaw, 2001; Holtom et al., 2008; Shaw, 2011). We also add greater specificity to the Job Embeddedness Model by exploring the network positions of individuals in the organization’s informal advice network and the required workflow network, and iden-tifying the positions that lead employees to harbor greater or lesser intentions to exit the organization. In terms of those advice-giving relationships that may increase turnover intentions, we explore the possibility that these behaviors are not adequately rewarded and test the mediating effect of employee perceptions of the fairness of rewards in the organization (distributive justice). We exam-ine these relationships in a fast-growing life sciences organization focused on food safety and animal health products driven by research performed by highly skilled knowl-edge workers. This organization’s continued rapid growth was threatened by a turnover rate over twice the industry average, motivat-ing them to participate in the study and to be concerned about their employees’ advice giving and seeking.

Social Network Infl uences on Turnover

The Job Embeddedness Model suggests that employee turnover is a function of a broad range of concerns, including ties to others, their job’s fit with other aspects of their life, and what they would need to sacrifice if they left (Mitchell et al., 2001). This study’s focus is on the ties—the relationships that an individ-ual has to other people or activities that embed them in the workplace—while controlling for fit and sacrifice. More specifically, we study those ties within the workplace that should have an influence on turnover intentions. While the full JEM includes ties within the workplace and those to family, community,

and other external entities, research suggests that those ties internal to the workplace have greater influence on attitudes such as satisfac-tion or commitment than do external ties (T. W. Lee, Mitchell, Sablynski, Burton, & Holtom, 2004). Also, to the extent to which fit and sacrifice are malleable employee attitudes, it stands to reason that ties, and especially in-ternal ties, would be the most crucial portion of the JEM warranting greater attention (Allen, 2006; cf. Felps et al., 2009).2

The few studies that focus on these ties using social network analysis typically take a social capital theory approach (e.g., Coleman, 1990), arguing that ties to other people within the organi-zation provide access to resources that make employees feel more “attached” to the organization. Studies using this social capital approach to turnover (Friedman & Holtom, 2002; Krackhardt & Porter, 1986; Mossholder, Settoon, & Henagan, 2005) have exam-ined different types of social ties between employees. The seminal study by Krackhardt and Porter (1986) explored communica-tion ties between employees at a fast-food restaurant. This study found that turnover occurred in clusters of employees who com-municated together intensively and who occupied similar struc-tural positions (patterns of ties to similar others). Mossholder et al. (2005) found that employees who were more central in informal net-works were less likely to turn over. They defined central individuals as those to whom others reported having either many social communication ties or work-related advice ties. Friedman and Holtom (2002) studied how membership in minority “network groups” provides social capital and support, which reduces turnover intentions.

Our study likewise explores how ties to employees affect an individual’s intention to turn over, but it differs from previous work in several key ways. Whereas Krackhardt and



Network theory and

research pushes

us to go beyond the

mechanism of “more

ties” to explain

embeddedness

to consider

combinations of ties

of a certain type

when explaining

phenomena.

Porter (1986) focused on the similarity of structural positions, we instead focus on the content of an employee’s ties (i.e., whether they are formally required by the organiza-tion or voluntary advice ties), as well as the extent to which those ties do or do not over-lap for individuals. Additionally, in contrast to Mossholder et al.’s (2005) research, we explore two specific types of work-related ties (required workflow and advice giving/seeking) to understand whether multiplex-ity in the relationship (the overlap or lack

of overlap between workflow and advice ties with the same cowork-ers) is related to turnover inten-tions. Our work also differs from the Friedman and Holtom (2002) study in that we are not examin-ing network or identity groups, but rather each individual’s per-sonal network as it is embedded in the overall organizational social network.

The manner in which the con-cept of links was examined ini-tially in the JEM has tended to be a bit vague in comparison to social network approaches. For example, Felps et al. (2009) operationalized it as the average embeddedness among an individual’s coworkers in a particular department, while Mitchell et al. (2001) asked indi-

viduals about their informal relationships in the workplace, including items such as “How many coworkers do you interact with regu-larly?” and “How many coworkers are highly dependent on you?” (Mitchell et al., 2001). However, without a social network perspec-tive, they were unable to get very specific on how, for instance, network relationships that others report to the focal individual are related to turnover. For example, previous JEM researchers have not asked which specific coworkers the respondent was tied to (needed to explore nominations by others) or about any other types of informal relationships beyond simple communication patterns that might also embed employees in the organiza-tion. The measure of links to the organization used in the JEM also included questions such

as “How long have you been in your present position?” “How long have you worked in the industry?” and “How many work teams are you on?” which are not specifically elicit-ing social ties.

In contrast, we focus on the specific social links to other employees within the organi-zation in order to isolate their effects (and we will thus employ the traditional social network reference to these links as “ties”). This level of specification enriches the JEM in several ways. First, it challenges the initial notion that more ties equal greater embed-dedness and, in turn, lower turnover inten-tions. Network theory and research pushes us to go beyond the mechanism of “more ties” to explain embeddedness to consider combinations of ties of a certain type when explaining phenomena. This is also in line with recent work by Oldroyd and Morris (2012), who suggest that having too much social capital (obtained through social ties) can become detrimental to an employee. Second, it delineates between multiple types of ties within the workplace accounting for both the informal emergent ties and the more formal, organizationally dictated ties. Finally, we specify two mechanisms by which social ties influence turnover intentions: directly, by providing social support, or indirectly, by influencing perceptions of distributive justice.

Two types of ties should theoretically influence the amount of social support an employee perceives or an employee’s percep-tions of distributive justice: those social ties mandated by the organization’s workflows and those that emerge informally through the seeking and providing of advice. We now turn to describing these relationships and the underlying manner in which they might be independently or jointly related to turnover intentions.

Mandatory Workplace Ties

In all organizations, employees experience some form of formal interdependence. A cur-sory glance at an organization chart can give some sense of how many other employees one has reporting to them and how many others one must report to. Going beyond the



organization chart, however, studies of work-flow networks directly ask employees ques-tions such as “whom are you required to work with to get your job done” and are therefore potentially able to capture both reporting lines and lateral task interdependencies (e.g., Brass, 1984; Umphress, Labianca, Brass, Kass, & Scholten, 2003). By examining this “work-flow network,” we can determine the extent to which individuals perceive that they are required to work with others and, conversely, how many others report a requirement to work with those individuals. This required workflow network reflects how employees in-terpret the collective desire of an organiza-tion’s managers and supervisors to direct the work-related interaction patterns within the workplace context.

Advice-Seeking Workplace Ties

No organization can operate only on the for-mally dictated ties and interactions between employees established by management, be-cause the environment with which the orga-nization deals is often more complex than the workflow network set up to handle it. Voluntary social ties, and particularly advice ties, are a critical source of flexibility and re-sources that allow tasks to be accomplished and organizations to adapt to changing cir-cumstances rapidly (e.g., Morey & Luthans, 1991). Contemporary organizations, particu-larly in knowledge-based industries, cannot operate effectively without this voluntary network of ties to help them adapt and main-tain their alignment with rapidly changing business environments. In turn, establishing and maintaining these voluntary social ties becomes an important employee capability needed to accomplish their tasks.

In their initial presentation of the JEM, Mitchell and colleagues (2001, p. 1104) sug-gest that “certain ties may be more important than others.” Advice-seeking relationships should be particularly important in deter-mining the level of support from the orga-nization’s social network experienced by the employee. Individuals actively choose whom they seek out for advice (rather than having it chosen for them by management) and so

these relationships should provide support that is strongly perceived and appreciated by the individual. Various streams of litera-ture speak to the importance of access (real or perceived) to others in the workplace to help create social capital (Coleman, 1990), social integration (O’Reilly, Caldwell, & Barnett, 1989), or general embeddedness (Mitchell et al., 2001), which all serve to further tie one to his or her organization. We expect that the more access an individual has to others for advice seeking, the less likely that individual will be to want to leave the organization.

Hypothesis 1: An employee’s number of advice-seeking ties will be negatively related to his or her turnover intentions.

To this point in this article, we have intro-duced the workflow and advice networks of the organization separately, which is typical of social network analysis. Network analysts often discuss and empirically test each type of rela-tion separately rather than combining or jux-taposing them, although this is clearly not how employees experience their social world in the workplace (e.g., Borgatti, Mehra, Brass, & Labianca, 2009). Increasingly, network researchers are considering how different types of relationships work together (LePine, Methot, Crawford, & Buckman, 2012; Shah, Bendersky, & Waldstrøm, 2011). In our con-text, it is important to consider how both advice giving and advice seeking (and pre-sumably receiving) might be quantitatively different when accompanied or devoid of workflow obligations with those same coworkers.

If an employee is able to seek out advice from coworkers who genuinely want to help, the employee is more likely to feel as if he or she is being done a favor and perceives greater social support (thus harboring lower turnover intentions). Conversely, if an employee seeks out advice from a coworker who is required to work with him/her, it may not be clear if the coworker is providing advice out of a desire to help or an obligation derived from their part of the organization’s workflow. This is not to say that there is no social support gained from hav-ing informal advice-seeking ties to those who



are required to work with the seeker. Rather, we argue that the employee might perceive that some of the advice may be given from a sense of obligation due to the required nature of their tie, thus conveying less social support that would serve to embed the employee in the organization. This idea is underscored in Figure 1a, where fictional Employee B seeks advice from two employees who are required to work with him or her and two other employees not required to work with him or her. Employee A, by contrast, is able to seek advice from all four others who are not obligated by workflow ties to provide assistance and is therefore more

clearly socially supported in the workplace (since those who are giving the advice are not being required to do so due to their perceived workflow obligations).

This idea is not entirely novel. The litera-ture surrounding extra-role or organizational citizenship behaviors (OCBs) is founded on the idea that such behaviors are beneficial when the behavior is discretionary and not directed by the organization (Organ, 1988). As helping behaviors such as advice giv-ing become part of the job, not only can conceptual and pragmatic confusion occur (Turnipseed & Wilson, 2009), but one would

(a) Advice-Seeking Ties

Circles represent fictional employees. A solid line represents an advice-seeking relationship. Adashed line represents a reported required workflow relationship. Both employees seek advicefrom four other employees (four advice-seeking ties). None of those whom employee A isseeking advice from report that they have to work with A. Two of those employee B reportsseeking advice from report that they have to work with B. Therefore, A would have four“voluntary advice-seeking ties,” while B would only have two.

(b) Advice-Giving Ties

Again, circles represent fictional employees. In this diagram, a solid line represents an advice-giving relationship. A dashed line represents a reported required workflow relationship. Both

employees give advice to four others (four advice-giving ties). Employee A does not report that

they are required to work with any of the four employees to whom they give advice. Employee B

reports that they are required to work with two of the employees who seek them out for advice.

Therefore, A would have no “required advice-giving ties,” while B would have two of thesemultiplex ties.

FIGURE 1. Advice and the Workfl ow Network



also expect the organizational benefits of OCB would likely diminish since it is no longer going “above and beyond.” Similarly, if one perceives that someone is giving them work-related help or advice due to a work obliga-tion, the positive psychological benefits of this action (including a potentially enhanced feeling of embeddedness) might be dimin-ished. Social support is most clearly conveyed to the recipient when employees are helped by others who have no obligation to do so.

Hypothesis 2: The number of advice-seeking ties to employees outside of the required workfl ow net-work (voluntary advice-seeking ties) will be nega-tively related to turnover intentions.

Advice-Giving Workplace Ties

One of the main ways in which the notion of social ties in the JEM needs greater nuance is in recognizing that each tie can be a source of both social capital as well as social liability—that is, every tie is not necessarily helpful for each employee (Labianca & Brass, 2006). The underlying assumption of the JEM is that in-creasing numbers of ties are positive and should decrease turnover intentions to vary-ing extents. However, it may be the case that some specific types of ties could actually lead to increased turnover intentions.

We argue that when employees are highly sought out for advice (that is, they are in the position of being advice givers), they are distracted by the request, must set aside what they are currently working on to consider the request, decide on whether to provide that assistance, actually provide the assistance or turn down the request for assistance, and then switch back to their own tasks (cf. Oldroyd & Morris, 2012). Research into task switching suggests that there is a steep cognitive price to pay when employees are forced to switch often (Mitchell et al., 2001; Waller, 2007). The employee who is in the situation of being asked advice by a large number of people would be more prone to burnout and may become frustrated with an inability to get his or her own work done due to the heightened demands of helping others.

Aside from the frustration that comes from switching tasks, employees who provide a great deal of advice also face cognitive bur-dens. When employees provide advice to one another, they must go through the process of interpreting the information embedded in the request, consider options to solve the problem, and choose and communicate what route they feel their coworker should take. As one can imagine, there is a great deal of infor-mation to be processed, suggesting the poten-tial for an “information overload” that might be detrimental (O’Reilly, 1980; Tushman & Nadler, 1978). Clearly, employees who have important skills will be sought out more than those without those skills. As is suggested by Oldroyd and Morris (2012), these “star employees” who have high human capital (i.e., knowledge, skills, abilities, and perfor-mance) will be more heavily sought out in organizational social networks, providing them with access to increased social capital. The downside of this situation is that these star employees might become overloaded with information that, in time, will hurt their performance, increase the likelihood of turn-over, and may even negatively impact organi-zational performance.

Hypothesis 3a: An employee’s number of advice-giving ties will be positively related to his or her turnover intentions.

The liabilities associated with advice giving would likely be enough to increase turnover intentions directly, but we further suspect that there will be an indirect effect through distributive justice. Advice giving or helping is typically more discreet and unlikely to be noticed by managers, which can lead to a lack of reward for advice giving. In stud-ies of how accurately individuals perceive the networks around them, managers have not fared well, with hierarchical position in the organization either being unrelated to accuracy (Krackhardt, 1990) or even nega-tively related (Casciaro, Carley, & Krackhardt, 1999). If managers not only frequently miss these behaviors, but also see (and potentially reward) advice and helping that does not actually exist, those employees who take the



time and effort to give advice are likely to feel the organization’s rewards are unjust. Equity theory (Adams, 1965) suggests that when employees are not being adequately rewarded for their efforts relative to others—that is, they perceive distributive injustice—they will take steps to create equity such as reducing effort, attempting to gain more rewards, or exiting the situation to search for something more equitable. With this backdrop, it is not sur-prising that extant research has shown strong negative effects of distributive justice on with-drawal behaviors and perceptions, suggesting those who perceive that the organization is rewarding unjustly will be more likely to har-bor increased turnover intentions (Colquitt, Conlon, Wesson, Porter, & Ng, 2001).

Hypothesis 3b: Distributive justice perceptions will mediate the relationship between an employ-ee’s number of advice-giving ties and his or her turnover intentions.

As was the case with advice seeking, we suspect that the intertwining of advice-giving and required workflow ties may again play an important role in revealing the employee’s turnover intentions. Whereas previously we sought to isolate advice seeking to those who are not required to work with the employee, as this would better represent social sup-port and reduce turnover intentions, for advice giving it is likely that where it overlaps with the required workflow ties, turnover intentions will increase. For example, if an employee were approached for advice by col-leagues he or she feels required to work with, the employee would have to oblige these information requests or risk upsetting the required workflow. These formally prescribed, intertwined workflow and advice-giving ties allow no escape from the task-switching costs of having to drop what you are working on to help the other person, and its attendant information overload, burnout, and frustra-tion. This hypothesis is illustrated in Figure 1b, where four others seek both employees out for advice. Employee A does not feel required to work with any of the four employ-ees who seek him or her out for advice, whereas Employee B would feel required to

give advice to two through the intertwined ties. These intertwined required advice-giving ties are therefore problematic for several rea-sons, as outlined earlier.

Hypothesis 4a: An employee’s number of advice-giving ties perceived to be required by the workfl ow (required advice-giving ties) will be positively re-lated to the employee’s turnover intentions.

While the advice giver might derive some intrinsic satisfaction from helping others, many organizations’ human resources systems are not designed to reward these types of behav-iors, even when these advice-giving behaviors might be very time-consuming and distract-ing from the employee’s primary role. Since many organizations preach collaboration but do not adequately reward collaborative behav-iors such as providing advice, we argue that those who give advice (particularly when they feel required to do so) will see the organiza-tion as low in distributive justice. Distributive justice exists when the goals of management or the organization are aligned with provided outcomes (Deutsch, 1975; Leventhal, 1976). When advice giving becomes perceived to be a required behavior (thus facilitating organiza-tional goals), intrinsic satisfaction is reduced and the need for extrinsic reward becomes greater in order to maintain a just organiza-tional rewards system. Given the propensity for advice-giving behaviors to be overlooked even if they are mandated by the organiza-tional workflow, distributive injustice is likely to increase with increased required advice-giving ties, and should act as a mediator lead-ing to increased turnover intentions.

Hypothesis 4b: Distributive justice will mediate the relationship between the employee’s number of required advice-giving ties and his or her turn-over intentions.

Method

Sample, Settings, and Procedure

The data for our study were collected from two divisions of a midsized food and animal



safety product manufacturing company lo-cated in two different states in the midwest-ern United States. A survey was administered including demographic, attitudinal, behav-ioral, and sociometric (network) items. There were 183 employees in Division One, of which 154 returned usable surveys (a re-sponse rate of 84.2 percent). This division also housed the corporate offices for the firm. The second division was a branch office in a differ-ent state that housed 99 employees, of which 75 participated fully (a rate of 75.8 percent). This gave us 229 usable surveys for a final overall response rate of 81.2 percent.

The employees were involved in a wide variety of jobs organized mainly in a tra-ditional functional structure that included research and development, manufacturing, warehousing, sales and marketing, and gen-eral administration. As a consequence, the organization employed individuals with a broad diversity of educational backgrounds, from high school degrees in areas such as warehousing to PhDs with significant research experience in their R&D department. The organization is a leader in a particular segment of its industry and was strategically pursuing numerous acquisitions in order to maintain its preeminent position, as well as to access the R&D and product capabilities of other smaller firms in its industry. This required it to maintain a tight control over its finances, which also meant that its pay levels were below the industry average, even within their geographic region. This contributed to a turnover rate that at the time of survey admin-istration was at roughly 22 percent, nearly twice the industry average (Culpepper Survey Trends, 2010). The HR systems were designed and maintained by the corporate HR depart-ment located in Division One, although there were two HR department members located within Division Two that assisted manage-ment and employees with HR-related issues, such as performance appraisals and schedul-ing. Employees were rewarded based on their manager’s annual performance assessment of their work as individuals, despite the fact that some departments worked interdependently to accomplish their tasks (e.g., R&D, produc-tion). The criteria for assessment were fairly

vague, and the assessments did not explicitly assess the employees’ interpersonal helping behaviors or other organizational citizenship behaviors.

Independent Variables

All of our independent variables were based on two sociometric questions. To create the workflow network, we presented a list of the names of all employees at the partici-pant’s division, along with the question “Are you required to work directly with this per-son in order to get your work done (e.g., re-ceiving inputs or providing outputs)?” All the data were input into UCINET VI (Borgatti, Everett, & Freeman, 2002). We aggregated these responses to create a matrix where a “1” exists in a cell if the name in that row re-ported that they feel as if they are required to work with the name in the column for each branch. Because we were interested in the perceived obligations that may exist when someone reports that they are required to work with someone else, we left our matrix unsymmetrized.

To capture the advice network, we posed the question, “Do you go to this person for work-related advice and knowledge?” fol-lowed by a list of names of all the employees in the division. One semantic detail of the wording of this item is of particular impor-tance for our setting. We asked employees who do they go to, rather than who would they go to, as is sometimes done when eliciting an advice network. By asking about the actual and not potential set of employees one seeks out, we have a higher level of certainty that advice is being sought out (and likely given). This is important because if one employee could seek out another (but never does), the employee being sought out would not have any of the frustrations that may come from dealing with requests for giving. We again aggregated these responses to create a matrix where a “1” exists in a cell if the name in that row reported that they go to the name in the column for advice. Advice seeking is an inherently asymmetric construct (i.e., we often seek advice from those who do not seek advice from us), and we were interested



theoretically in understanding the distinction between advice seeking and advice giving, so we left this matrix unsymmetrized as well. To construct our measures of advice seeking, we summed each individual’s row of the matrix that counts how many others each employee said they go to for advice (technically known as outdegree centrality). For advice giving, we summed each individual’s column that indi-cates the number of times someone was listed by another employee as someone whom they would seek out for advice (also known as indegree centrality).

To obtain the number of times an advice-seeking tie was not accompanied by a required workflow tie, we combined the two network matrices. Here we were interested in ensuring that the advice-seeking tie was not embedded in a required incoming workflow relationship. We first transposed the workflow network so that a “1” in a row would now indicate that the person in the column reported that they are required to work with the person in the row. We then recoded the advice network so that an advice tie would be represented by a “2” instead of a “1.” Finally, we added the transposed workflow network to the recoded advice network so that a value of “1” in a row would represent the presence of an incoming workflow-only tie, a “2” would represent an advice-seeking-only relationship, and a “3” would indicate that there is an intertwined advice-seeking and a required workflow rela-tionship reported by the advice seeker. We counted all “2s” in each row as the number of advice-seeking ties without any required workflow requirements from the individual giving help and advice. We labeled these advice-seeking ties to employees not required to work with the seeker “voluntary advice-seeking ties” in the tables.

The procedure for computing the presence or absence of advice-giving ties intertwined with required workflow ties again involved combining these two matrices. To measure the number of intertwined required workflow outgoing ties and incoming advice-giving ties overlaps, we employed matrix algebra. We transposed the advice network so that rows would now represent the individuals who were seeking advice from the focal employee.

We then added the workflow matrix to the transposed advice network so that a value of “2” in a cell would indicate that the employee listed in the row is both required to work with and is sought out for advice from the employee in the column. We then counted the number of “2s” in each row to act as our measure of advice-giving ties intertwined with a required workflow relationship, and we labeled these “required advice-giving ties.”

Our mediating variable, distributive jus-tice, was adapted from a scale developed by Leventhal (1976). The three items included in our survey were: “Do organizational rewards reflect the effort you have put into your work?” “Are organizational rewards appropri-ate for the work you have completed?” and “Do organizational rewards reflect what you have contributed to the organization?” The scale had a Cronbach’s alpha of .87.

Dependent Variable

We employed the three-item Intention to Turnover scale from the Michigan Organizational Assessment Questionnaire (Cammann, Fichman, Jenkins, & Klesh, 1979). The scale used the following items: “I often think about quitting,” “I will probably look for a new job in the next year,” and “How likely is it that you could find a job with another employer with about the same pay and benefits you now have?” The scale had a Cronbach’s alpha of .66.

Controls

Although the primary focus of this study is on the “links” portion of the Job Embeddedness Model, we control for fit by including indi-viduals’ organizational tenure in our analyses. Those who have been in the organization for a longer period should possess greater fit be-cause they would otherwise likely have already exited the organization either voluntarily or involuntarily (Schneider, 1987). Tenure has been shown to have a positive relationship to several types of person–environment fit in meta-analytic research (Kristof-Brown, Zimmerman, & Johnson, 2005). We statisti-cally control for the “sacrifice” portion of the



Those who have

been in the

organization for a

longer period should

possess greater fit

because they would

otherwise likely

have already exited

the organization

either voluntarily or

involuntarily.

JEM by accounting for continuance organiza-tional commitment (Cronbach’s alpha = .55; Meyer & Allen, 1984).

We included a series of additional demo-graphic, attitudinal, and behavioral controls in our analyses that have been found to be important in meta-analytic work on turnover (Griffeth et al., 2000) and turnover intentions (Meyer, Stanley, Herscovitch, & Topolnytsky, 2002; Tett & Meyer, 1993). These included the demographic variables of gender, age, and full- or part-time status, and whether the individual was in a supervisory role. We also controlled for the individual’s human capital by including the employee’s level of educa-tion (which in this organization ranged from a high school education to a PhD), since this might affect advice giving and seeking, as well as external job prospects. We also controlled for the division in which the individual was employed, as the difference in the divi-sions’ sizes could have affected our network measures.

We controlled for several attitudinal mea-sures that have been shown to be important in predicting turnover intentions. We included a three-item job-satisfaction scale (Cammann et al., 1979; Cronbach’s alpha = .89) and a four-item affective commitment derived from Meyer and Allen (1984) with a Cronbach’s alpha = .80. Finally, we included a four-item self-reported measure of interpersonal organi-zational citizenship behaviors (OCBs) based on K. Lee and Allen (2002) to act as a proxy for how much time an individual spends at work helping others in order to isolate the effects of advice giving and receiving more specifically. The items used in our survey were “Help others who have been absent,” “Willingly give your time to help others who have work-related problems,” “Adjust your work schedule to accommodate other employees’ requests for time off,” and “Go out of the way to help newer employees feel welcome in the work group.” The OCBI scale’s Cronbach’s alpha was .71.

Results

Table I presents the means, standard deviations, and correlations among all variables. As ex-pected, there were strong correlations between

our measures of advice seeking/giving and workflow/advice overlaps, with some correla-tions reaching as high as 0.87. This created some multicollinearity concerns in subsequent regression models. In particular, attempting to place the advice giving and advice seeking in the same model as the multiplex measures cre-ated variance inflation factors (VIFs) over 4 in some models and over 8 in others, which were unacceptably high. These issues stem from the large degree of overlap between the two net-works and the fact that the multiplex measures are actually subsets of the advice-giving and advice-receiving measures (see the next section, for more details; full models, including all the VIFs are available upon request). We dealt with this by running separate models for hypotheses involving overall numbers of advice-seeking and advice-giving ties and for our two multiplex measures.3 There were also strong negative cor-relations between the traditional psychometric measures of organiza-tional withdrawal and the inten-tion to turnover (−0.67 and −0.68 for job satisfaction and affective organizational commitment, respec-tively), as well as between distribu-tive justice and turnover intention. Among the controls, age and tenure were both significantly negatively correlated to turnover intentions.

We used stepwise ordinary least squares (OLS) regression analyses to test our hypotheses. Our analy-ses included five models. Model 1 included all of our demographic, attitudinal, and behavioral con-trols. Model 2 included our advice-seeking and advice-giving measures. Model 3 added our mediator, distributive justice. Model 4 included only our controls and our two measures of workflow/advice multiplex ties (voluntary advice-seeking and required advice-giving ties). Finally, Model 5 adds the mediator, distributive justice, to the variables found in Model 4.

Advice-Seeking Workplace Ties

We expected that a higher number of advice-seeking ties would be negatively related to



TA

BL

E

I

Me

an

s, S

tan

da

rd D

evi

ati

on

s, a

nd

Co

rre

lati

on

s

M

ean

SD

12

34

56

78

910

1112

1314

1516

1.

In

ten

t to

Tu

rn o

ver

2.9

7.9

0

2. D

ivis

ion

1.33

.47

.09

3. S

up

erv

iso

ry

Ro

le.2

3.4

2−

.10

.07

4. F

ull-T

ime/

Part

-Tim

e

Sta

tus

.96

.19

.07

.09

.11

5. G

en

der

.51

.50

−.0

4.0

4.1

6*

.16*

6. A

ge

37.

82

11.4

1−

.17**

−.1

2.1

8**

.05

.14*

7.

Ten

ure

4.0

44.2

2−

.14*

−.0

2.3

4***

.04

.12

.39***

8. E

du

cati

on

4.1

51.

37

.07

−.0

6.2

2***

−.0

4.0

9−

.05

.12

9. Jo

b

Sati

sfa

cti

on

3.8

4.8

9−

.67***

−.0

5.0

9−

.16**

.01

.09

.12

−.0

8

10. A

ffecti

ve

Org

an

izati

on

al

Co

mm

itm

en

t

2.9

6.7

8−

.68***

−.0

5.2

0***

−.1

2.0

6.1

1.2

2***

−.0

3.7

0***

11.

Co

nti

nu

an

ce

Org

an

izati

on

al

Co

mm

itm

en

t

3.0

6.5

8−

.47***

−.1

4*

.01

−.0

9−

.03

.16**

.22***

−.0

8.4

0***

.51

**

*

12. I

nte

rpers

on

al

Cit

izen

sh

ip

Beh

avio

rs

3.8

3.8

1−

.06

.08

.12

.06

−.0

9−

.05

.00

−.1

2.1

7**

.15

*.0

0

13. A

dvic

e-

Seekin

g T

ies

4.8

84.1

9−

.08

.24***

.15*

.10

.01

−.0

8.0

4.1

6*

.05

.12

.12

.15

*



14. A

dvic

e-G

ivin

g

Tie

s4.8

84.1

9.0

2.2

4***

.62***

.16*

.01

.06

.35***

.26***

.05

.14

*.0

2.1

1.2

9*

**

15

. Req

uir

ed

Ad

vic

e-G

ivin

g

Tie

s

2.3

12.4

5.0

5.2

4***

.49***

.13

−.0

4.0

4.2

5***

.20***

.05

.14

*.0

3.2

0*

**

.49

**

*.7

4*

**

16. V

olu

nta

ry

Ad

vic

e-

Seekin

g T

ies

2.5

72.8

7−

.10

.15*

.00

.05

.06

−.1

3−

.06

.14*

.06

.05

.08

.14

*.8

7*

**

.02

.24

**

*

17.

Dis

trib

uti

ve

Ju

sti

ce

2.4

9.9

7−

.42

.11

.12

−.0

5.0

4−

.02

−.0

5−

.07

.40

.43

.19

−.0

1.0

0−

.07

.00

−.0

9

*C

oeffi

cie

nt

is s

ign

ifi c

an

t at

the .05 level (t

wo

-tailed

).

**C

oeffi

cie

nt

is s

ign

ifi c

an

t at

the .01 level (t

wo

-tailed

).

***C

oeffi

cie

nt

is s

ign

ifi c

an

t at

the .0

01 level (t

wo

-tailed

).



While the overall

number of

coworkers an

employee seeks

out for advice does

not impact turnover

intentions, the ability

to seek advice from

coworkers outside

of regular required

workflow ties

decreases turnover

intentions.

turnover intentions because employees who could reach out to get the advice they needed would be less motivated to leave the organi-zation (H1). However, as seen in Model 2 of Table II, there was no significant relationship between advice seeking and turnover inten-tions (b = −.04, p > .10). We did, however, find support for Hypothesis 2, which stated that when an employee is able to get advice from someone who is not required to help, they will be less likely to have turnover inten-tions (Model 4, b = −.12, p < .05). When these results are taken together, we find that while the overall number of coworkers an employee seeks out for advice does not im-pact turnover intentions, the ability to seek

advice from coworkers outside of regular required workflow ties de-creases turnover intentions. This suggests that other employees’ willingness to engage in purely voluntary helping behaviors (i.e., advice giving) helps bind the focal employee to the organization.

Advice-Giving Workplace Ties

Hypothesis 3a exploring advice giving was supported, as those employees who were heavily sought out for advice were signifi-cantly more likely to have higher turnover intentions (Model 2, b = .15, p < .05). We found similar support for Hypothesis 4a, which suggested that when employees were sought out for advice and these advice ties were intertwined with required workflow relation-ships, turnover intentions would be higher. This was, indeed, the

significant relationship we found between re-quired advice-giving ties and intent to turn-over (Model 4, b = .19, p < .01).

Across these stepwise models, we see that the control variables as a whole explain 56.6 percent of the variance in turnover inten-tions. Adding network variables directly to these models explains up to 3.8 percent of additional variance in the dependent variable.

All of the results were replicated (albeit with slightly weaker levels of significance) if we used a required workflow network symme-trized on minimum, where both respondents reciprocated a required workflow tie (results available on request from the authors).

If we remove the traditional attitudinal psychometric controls for job satisfaction and organizational commitment, all of which were highly correlated with turnover inten-tions (r from 0.47–0.68) and were collected at the same time as the turnover intentions, and thus potentially most vulnerable to percept-percept bias (Crampton & Wagner, 1994), the variance explained in the control model drops to 7 percent. Including the network variables in this control model doubles the variance explained to 14 percent (results not shown). We should note, however, that social network measures such as advice seeking might also be somewhat susceptible to percept-percept bias. However, our self-reported measure of advice seeking was not significantly related to turnover intentions, suggesting that this type of bias was less of an issue in our case.

Distributive Justice as a Mediator

We also argued that distributive justice was the mediating factor between advice giving and turnover intentions. As employees take on a greater advice-giving role in the organi-zation, we argued that they would feel that their rewards were not keeping up with the additional effort they were putting in, and their turnover intentions should increase. We tested for the mediating effect of distributive justice on the relationship between the em-ployee’s number of advice-giving ties and turnover intentions using the significance of indirect effects approach (MacKinnon, Fairchild, & Fritz, 2007; Preacher & Hayes, 2004; Shrout & Bolger, 2002). We used both the Sobel test (Sobel, 1982) and bootstrapped confidence intervals (1,000 permutations) around the indirect effect. We found that dis-tributive justice was negatively related to turn-over intentions (Figure 2a, Path B: b = −.14, p < .05). Furthermore, the indirect effect of advice-giving ties on turnover intentions through distributive justice was significant



T A B L E I I Results of a Hierarchical OLS Regression Analysis

Model 1 Model 2 Model 3 Model 4 Model 5

(Constant)

Branch.052

(.088)

.010

(.093)

.054

(.094)

.005

(.089)

.042

(.089)

Supervisory Role−.008

(.109)

−.174

(.129)

−.107

(.131)

−.181

(.116)

−.133

(.117)

Full-Time/Part-Time

Status

−.227

(.216)

−.282

(.216)

−.257

(.214)

−.276

(.211)

−.258

(.209)

Gender−.010

(.084)

.023

(.084)

.016

(.083)

.053

(.083)

.046

(.082)

Age−.006

(.004)

−.006

(.004)

−.006

(.004)

−.007

(.004)

−.007

(.004)

Tenure.009

(.011)

.003

(.011)

.000

(.011)

.004

(.011)

.000

(.011)

Education.013

(.031)

.004

(.032)

.003

(.032)

.009

(.031)

.007

(.031)

Job Satisfaction−.384***

(.065)

−.385***

(.064)

−.360***

(.064)

−.371***

(.063)

−.349***

(.063)

Affective Organizational

Commitment

−.411***

(.080)

−.411***

(.080)

−.357***

(.082)

−.430***

(.078)

−.380***

(.081)

Continuance Organiza-

tional Commitment

−.217*

(.084)

−.212*

(.085)

−.214*

(.084)

−.199**

(.083)

−.203**

(.082)

Interpersonal

Citizenship Behaviors

.065

(.052)

.067

(.052)

.053

(.052)

.058

(.052)

.046

(.052)

Advice-Seeking Ties−.009

(.010)

−.010

(.010)

Advice-Giving Ties.033*

(.014)

.026

(.014)

Voluntary Advice-

Seeking Ties

−.036*

(.015)

−.036*

(.015)

Required Advice-Giving

Ties

.070***

(.020)

.061**

(.021)

Distributive Justice−.115*

(.049)

−.106*

(.048)

Constant 6.382 6.419 6.488 6.477 6.535

R-Squared .566 .578 .589 .595 .604

Change in R2 .566 .012 .008 .028 .009

F for Change in R2 25.63*** 3.09* 2.029 7.522*** 4.894*

Unstandardized coeffi cients reported above with standard errors in parentheses.

*Coeffi cient is signifi cant at the .05 level (two-tailed).

**Coeffi cient is signifi cant at the .01 level (two-tailed).

***Coeffi cient is signifi cant at the .001 level (two-tailed).



(Sobel z = 1.92, p < .05; 95 percent CI around indirect effect: [.0009, .0172]). Our analyses demonstrated that when distributive justice is accounted for as a mediator, advice giving no longer had a significant effect on turnover in-tentions (Figure 2a, Path C’), suggesting a full mediation effect in support of Hypothesis 3b.

We performed similar tests for the medi-ating effect of distributive justice on the rela-tionship between advice-giving ties that were intertwined with required workflow relation-ships and turnover intentions. We found a significant, negative relationship between the employee’s number of required advice ties and his or her perceptions of distributive justice (Figure 2b, Path A: b = −.23, p < .01). Given the previously established negative relation-ship between distributive justice and turnover intentions, we then tested the indirect effect of advice-giving ties to required work partners.

The indirect effect of these particular advice-giving ties on turnover intentions through dis-tributive justice was significant (Sobel z = 1.83, p < .05; 95% CI around indirect effect: [.0009, .0251]). Our results suggest a partial mediation effect as the number of required advice-giving ties is still significantly related to intent to turn over after distributive justice is accounted for, albeit to a lesser extent (Figure 2b, Paths C and C’), supporting Hypothesis 4b.

In sum, these findings suggest that those who are heavily sought out for advice become overextended, do not perceive themselves to be adequately rewarded by the organization for their extra efforts, and are more likely to intend to leave. This is especially acute when employees are obliged to honor those advice requests because they are coming from employ-ees whom managers are requiring them to work with due to workflow considerations.

(a) Advice Giving

(b) Required Advice Giving

Unstandardized coefficients reported above with standard errors in parentheses.

*Coefficient is significant at the .05 level (two-tailed).

**Coefficient is significant at the .01 level (two-tailed).

***Coefficient is significant at the .001 level (two-tailed).

FIGURE 2. Mediation Results



Our main theoretical

contribution is

to clarify the

relationship between

social ties and

turnover intentions

suggested in the

Job Embeddedness

Model by

demonstrating that

the mere presence

of ties in the

workplace will not

necessarily deter

employees from

considering leaving

the organization.

Discussion

Our study contributes to current theorizing regarding the turnover process and opens the door to a multitude of possibilities for better integrating social network analysis into stud-ies of turnover intentions. Our main theoreti-cal contribution is to clarify the relationship between social ties and turnover intentions suggested in the Job Embeddedness Model by demonstrating that the mere presence of ties in the workplace will not necessarily deter employees from considering leaving the orga-nization. Rather, some types of ties were ben-eficial for keeping employees from strongly considering quitting, such as when employ-ees were able to seek advice from others who are not required to work with them. However, we also found situations where too many of certain types of ties actually increase intention to quit. For example, when an employee is being sought out by many others for work-related advice, the employee’s turnover in-tentions rise significantly. This underscores the importance of social ties and demon-strates that operationalizing these ties in a so-ciometric, network fashion as opposed to the Likert-type scales typically used in JEM stud-ies can provide a great deal more understand-ing of what leads to turnover intentions.

Another theoretical contribution we offer is in showing the mediating role of distributive justice perceptions. Those employees who are heavily taxed in the informal advice-giving network feel as if they are not being rewarded adequately for these behaviors, as captured by their lower distributive justice scores, which increased their desire to exit the organiza-tion. We also found that when employees are informally approached for advice by those whom they would be obligated to help due to mandatory workflow requirements, they are more likely to intend to turn over.4 This result was only partially mediated by distribu-tive justice, suggesting either a direct effect or the potential for these ties to influence other mediators not included in this study. Future research on the effects of social ties in the JEM should be careful to take a nuanced view of both the content of ties and the social context within which they exist.

Implications for Practice

Our study also provides practical insights to managers who are concerned about turnover. While there is a difference between intent to turnover (what we studied) and actual turn-over, meta-analytic research (e.g., Griffeth et al., 2000) has established a clear linkage be-tween intention and behavior suggesting managers should pay attention to this intent. One need not know the intricacies of the in-formal social structure to be able to imple-ment changes that attempt to accentuate the network features that reduce turnover inten-tions and reduce those features that lead to increased intent to quit. An important step management can take is to at-tempt to cultivate extra-workflow informal ties. Employees like being able to get advice from those outside of their formal workflows and do not like when they are forced to give advice to those with whom they work directly. Research by Kaše, Paauwe, and Zupan (2009) suggests that HR practices are able to exert at least some in-fluence on the social structures within organizations where knowledge management is crucial (as is the case in our setting). Facilitating interaction between employees who may not be re-quired to interact could be a place to start trying to create more ro-bust social structures. Establishing many types of informal workplace groups could be fertile ground for the cultivation of extra-workflow informal ties stretching beyond the coworkers with whom they are required to interact (cf. Friedman & Holtom, 2002). Organizational actions such as creating communities of practice, establishing online-mediated knowledge sharing, utilizing men-toring programs, and cross-staffing informal committees are some potential ways to aid in the creation of more robust informal networks.



If managers are

cognizant of the

efforts of their

employees in

helping coworkers

and can provide

either formal (pay)

or informal (praise)

rewards for those

who are particularly

helpful to their

coworkers, the

effect on turnover

intentions may be

diminished or even

reversed.

Another finding that should be of par-ticular interest to managers is that those who are frequently sought out for advice are more likely to consider leaving the organiza-tion. The frustration associated with required advice giving may push key employees out of the organization, and it may also be the case that this prominence within the orga-nization can also lead to forces pulling the employee out of the organization (i.e., other firms actively recruiting these employees). Whether it is frustration pushing the individ-uals toward leaving or social capital creating a market pulling the employee from the organi-zation, the fact that they are considering exit-ing the organization is troubling due to the

losses in both human and social capital. First, if these employees do leave the organization, they take with them their knowledge, which is what made them so sought out by coworkers in the first place. This loss of human capital could be very detrimen-tal to organizational functioning. Additionally, when these employ-ees leave, they leave a gap in the social capital of the organization. Advice-giving/-seeking relation-ships are not prescribed by the organization, but rather emerge informally over time. When a gap is created in the advice network, it takes time to fill and this could create a reduction in the organiza-tion’s efficiency and effectiveness in the interim period.

Ideas for how to avoid these gaps can be seen when exploring ways to reduce information over-load. To prevent star employees from burning out when heav-ily sought out for information, Oldroyd and Morris (2012) sug-gest organizations take steps to make information search pro-cesses more efficient by mak-

ing the right information easier to access via training or information technology. We suggest that these tactics could also reduce the amount of advice needed to be given

within the organization and would therefore weaken the relationship between advice-giving and turnover intentions. Furthermore, if employees still exit the organization after these tactics have been implemented, the det-riment should be lessened, as there would be systems in place for new employees to more easily utilize (i.e., an online knowledge repos-itory or a listing of who to call when various problems arise).

Another way to eliminate many of the detriments that emerge from employees being overutilized in the advice network may be to ensure that these behaviors are being adequately monitored and rewarded. Recall that distributive justice mediated the effect of advice giving on turnover intentions. Those giving a great deal of advice did not feel that the organization fairly rewarded their efforts and were more likely to harbor higher turn-over intentions. If managers are cognizant of the efforts of their employees in helping coworkers and can provide either formal (pay) or informal (praise) rewards for those who are particularly helpful to their cowork-ers, the effect on turnover intentions may be diminished or even reversed. This particular organization had no mechanism in place to appraise these behaviors. Potential options to increase managers’ awareness include instituting 360-degree feedback so that they can hear when employees are being helpful in the organization’s social networks and pro-viding managerial training on the effects of employee perceptions of justice (Greenberg, 2004).

Limitations and Future Research

Our results suggest that simply looking at the reports of key informants on how many oth-ers they go to for advice or taking a cursory glance of organization or process charts might not prove overly enlightening in predicting who is likely to leave. We used in-depth social network analysis to enhance our understand-ing of the effects of various types of social ties in the workplace. Of the two most straightfor-ward hypotheses regarding advice seeking and giving, we found that only one (advice giving) had a relationship with intent to turnover, a



Replication of our

results in other

settings, possibly

in a very different

context in which

such behaviors

are formally

acknowledged and

appraised, would

greatly bolster the

generalizability of

our findings.

relationship that was fully mediated by per-ceptions of distributive justice. It appears that even the significant finding for advice giving only tells part of the story given the complex nature of these relationships. These relation-ships may be moderated by factors currently not included in the research design such as characteristics of those who are giving or re-ceiving the advice (e.g., whether they are a su-perior, a subordinate, or a coworker with a similar job). In addition, while we find partial mediation for this result through distributive justice, other factors, such as advice quality, frequency, and complexity, may alter the na-ture of this relationship, while characteristics of those involved in the specific dyads may alter the strength of these effects. Testing these possibilities would require a more elaborate research design and provide suggestions for future research endeavors in the area of social networks with potential impact in the HRM field.

Our study was cross-sectional in nature and did not allow us to investigate changes in social ties and implications of such churn over time (Oldroyd & Morris, 2012; Sasovova, Mehra, Borgatti, & Schippers, 2010). For instance, we assume that advice-seeking behavior results in an actual advice flow. Perhaps part of network churn in the advice ties (lost advice-seeking ties) can be related to the fact that the person’s problems were not adequately solved or the requests were repeat-edly turned down, which may be a reason for turnover. On the other hand, a different part of network churn (new advice-giving ties) may be an indication of an increasing burden on the advisor and, possibly with some delay, lead to an increase in the person’s social sta-tus within the organization.

We also had concerns about the low reliability of our measure of the control variable continuance organizational com-mitment (Cronbach’s alpha = .55). These concerns were partly alleviated due to the fact that we used a widely validated measure developed by Meyer and Allen (1984) and that when rerunning the analyses without this control variable, our pattern of results remained unchanged. Moreover, a relatively low reliability usually means that it becomes

increasingly difficult to find any significant relations with other variables, but we find consistent significant effects of this control variable in the predicted direction.

A final limitation is related to the fact thatwe utilized a single sample in our study that raises concerns about the generalizabil-ity of our findings. While we only focused on one organization, the organization we stud-ied contained a diverse array of jobs, ranging from janitorial employees to administrative support staff to research scientists, suggest-ing some generalizability to our findings. Furthermore, our findings that advice giving increases turnover intentions center on the idea that these behaviors were overlooked and not adequately rewarded by the organization (hence, the mediation through distributive justice). We suspect that since these informal helping behaviors have been shown to be difficult to accurately observe (Casciaro et al., 1999; Krackhardt, 1990), this situation repeats itself in many organizations. Still, replication of our results in other settings, pos-sibly in a very different context in which such behaviors are formally acknowledged and appraised, would greatly bolster the general-izability of our findings.

Our study suggests strongly that the type of social tie (or combination of ties) studied is very important in determining attitudes regarding quitting. We looked at two types of instrumen-tal ties (required workflow ties and advice ties), but this is far from an exhaustive list of the types of ties that might affect turnover intentions. Negative ties (such as interpersonal dislike or conflict) could play a very important role in pushing individuals out of the organization (Labianca & Brass, 2006). Recently, a social network perspective has been applied to positive and negative gossip in organizations (Ellwardt, Labianca, & Wittek, 2012; Grosser, Lopez-Kidwell, & Labianca, 2010), and these studies suggest that particular individuals, including



A longitudinal study

that integrates the

concept of shocks

with a social

network perspective

on turnover

intentions could

prove illuminating.

those who are socially isolated, might be tar-geted for negative gossip by other employees, which might increase turnover intentions. Building on these perspectives by examining the targets of negative gossip and the social structures within which they are embedded could also provide new insights into employ-ees’ turnover intentions.

Yet another network lens that could be fruitfully applied to the study of turnover, and that would further elaborate the JEM, would be to extend Krackhardt and Porter’s work (1986) on individual perceptions of networks within organizations. Other work on the cog-nition of networks has shown that even per-ceived ties could be real in their consequences

(Kilduff & Krackhardt, 1994). It could well be the case that where people perceive themselves as fit-ting into the social structures of the organization could affect their turnover intentions. This idea would be commensurate with the fit dimension of the JEM, since Mitchell and colleagues (2001, p. 1104) define fit as “an employee’s perceived compatibility or comfort with an organization and with his or her environment” (italics added).

While we utilized the JEM as our primary theoretical lens, our research could also be extended

to contribute to the unfolding model of turn-over (T. W. Lee & Mitchell, 1994). The unfold-ing model of turnover suggests that turnover frequently begins with some sort of shock that causes an employee to reevaluate his or her situation. After the shock, the employee begins to consider exiting the organization and takes stock of his or her satisfaction, fit with the organization, and job alternatives. Employees who have a strong social support system within the organization may be insu-lated from shocks or, when a shock occurs, evaluate their overall situation favorably, both of which result in reduced turnover intentions. Additionally, information fre-quently travels through informal channels, and so employees who occupy different posi-tions in the informal networks of the orga-nization would have more or less access to

information about shocks that may, in turn, influence how the shock information is pro-cessed. A longitudinal study that integrates the concept of shocks with a social network perspective on turnover intentions could prove illuminating.

In conclusion, informal social structures exist beyond the formal workflow require-ments of a job and play a large role in deter-mining how work gets done (Krackhardt & Hanson, 1993). The human resource man-agement community could be strengthened in research and in practice by better account-ing for the informal networks of organiza-tions. Our study spoke to the importance of these ties in networks on the issue of turn-over that has been of theoretical and practical significance since the dawn of research and practice in human resource management. The research presented here elaborates the Job Embeddedness Model and shows that the complex interplay between formal and informal social ties influences employee pro-pensity to harbor turnover intentions. Given what we have learned over the decades about the diffusion of attitudes and practices within and between organizations, antecedents to employee performance, and the general complexity of organizational life by employ-ing a social network perspective, we can see that the potential to turn these volumes of research toward HR-specific issues is virtually boundless.

Notes1. We acknowledge the “Unfolding Model of Turn-

over” initially developed by T. W. Lee and Mitchell

(1994) offers an alternative perspective on the turn-

over process and discuss implications of our re-

search for this line of inquiry in the discussion

section.

2. Although the primary focus of this study is on the

“links” portion of the JEM, in our analyses we sta-

tistically control for the “sacrifice” portion of the

Model by accounting for continuance organiza-

tional commitment (Meyer & Allen, 1984). Also, we

control for tenure, which if applying Schneider’s

(1987) Attraction-Selection-Attrition model would

constitute a form of “fit” since those who have

been in the organization for longer times should

possess greater fit; otherwise they would have



likely exited the organization either voluntarily or

involuntarily. Tenure has been shown to have a

positive relationship to several types of person–

environment fit in meta-analytic research (Kristof-

Brown et al., 2005).

3. An alternative way to deal with the multicollinear-

ity issue would have been to construct two addi-

tional multiplex measures (required advice-seeking

ties and voluntary advice-giving ties). We con-

ducted these analyses (available upon request) and

found that our pattern of results was not signifi-

cantly changed.

4. We recognize that these findings are somewhat

contradictory of the findings of Mossholder et al.

(2005), but this is likely due to a key difference be-

tween that work and the present inquiry. While the

network centrality index used in their prior work

was measuring indegree, it combined social (com-

munication) and work-related (advice) links,

whereas we look solely at work-related ties and

examine networks separately and in combination.

Also, Mossholder and colleagues were looking at

actual turnover, whereas we are exploring turnover

intentions. It could well be the case that frustration

with unrewarded behaviors in the work-related

network can be overcome by strong social rela-

tions when it comes to the actual quitting decision.

In addition, direct or indirect social ties can help

create interpersonal obligations and a better lever-

age for dealing with work-related resource depen-

dence (Gargiulo, 1993; Gargiulo, Ertug, & Galunic,

2009).

SCOTT M. SOLTIS (PhD, University of Kentucky) is an assistant professor of management

at the University of Missouri-St. Louis. His primary research focus is at the crossroads

of human resource management and social network analysis, applying a social net-

work perspective to topics such as turnover, employee transfers, recruitment, perform-

ance, and person–environment fi t. His research has been published in Organizational

Psychology Review and the American Psychological Association Handbook of Industrial

and Organizational Psychology.

FILIP AGNEESSENS is a researcher in the Department of Sociology/ICS at the University

of Groningen. He has published a number of articles on social network analysis in lead-

ing social science journals. His research primarily focuses on network formation within

organizations and their impact on attitudes and behavior of individual employees. He

has also been working on the development of methods for analyzing networks, imputa-

tion techniques for solving missing data problems, and methods for building social sup-

port typologies for ego networks. He has taught numerous courses on social network

analysis and organizational research.

ZUZANA SASOVOVA is an assistant professor in the Faculty of Economics and Business

Administration affi liated with an interdisciplinary Network Institute at the VU University

in Amsterdam. She received her PhD in information systems and organizational behav-

ior in 2006. Her recent work has been published in Administrative Science Quarterly and

the Journal of Product Innovation Management. Her research interests include dynamics

of social networks, cognitions, and social processes of innovation.

GIUSEPPE (JOE) LABIANCA (PhD, Penn State) is a Gatton Endowed Professor of

Management at the University of Kentucky and co-director of the LINKS Social Network

Analysis Center (linkscenter.org). His research takes a social network perspective on

organizational behavior topics such as groups, power, confl ict, and gossip. His work

has appeared in Science, the Academy of Management Journal, the Academy of

Management Review, Harvard Business Review, Organization Science, Sociology

of Organizations, Social Networks, Management Information Science Quarterly, and else-

where. He serves on the editorial boards of Organization Science, Strategic Organization,

and Organizational Psychology Review.



ReferencesAdams, S. J. (1965). Inequity in social exchange. In

L. Berkowitz (Ed.), Advances in experimental social

psychology (Vol. 2, pp. 267–299). New York, NY:

Academic Press.

Allen, D. G. (2006). Do organizational socialization

tactics infl uence newcomer embeddedness and

turnover? Journal of Management, 32, 237–256.

Borgatti, S., Everett, M., & Freeman, L. (2002). UCINET

for Windows: Software for social network analysis.

Lexington, KY: Analytic Technologies.

Borgatti, S. P., Mehra, A., Brass, D. J., & Labianca, G.

(2009). Network analysis in the social sciences.

Science, 323, 892–895.

Brass, D. (1984). Being in the right place: A structural

analysis of individual infl uence in an organization.

Administrative Science Quarterly, 29, 518–539.

Cammann, C., Fichman, M., Jenkins, D., & Klesh,

J. (1979). The Michigan organizational assess-

ment questionnaire (Unpublished manuscript).

University of Michigan, Ann Arbor, MI.

Casciaro, T., Carley, K. M., & Krackhardt, D. (1999).

Positive affectivity and accuracy in social network

perception. Motivation and Emotion, 23, 285–306.

Coleman, J. S. (1990). Foundations of social theory.

Cambridge, MA: Belknap Press of Harvard

University Press.

Colquitt, J. A., Conlon, D. E., Wesson, M. J., Porter, C.

O., & Ng, K. Y. (2001). Justice at the millennium: A

metaanalytic review of 25 years of organizational

justice research. Journal of Applied Psychology, 86,

425–445.

Crampton, S., & Wagner, J. (1994). Percept-percept in-

fl ation in microorganizational research: An investi-

gation of prevalence and effect. Journal of Applied

Psychology, 79, 67–76.

Culpepper Survey Trends. (2010). Turnover: A two-

edged sword that must be watched closely

Retrieved from http://www.culpepper.com

/eBulletin/2010/Turnover.asp

Dess, G., & Shaw, J. (2001). Voluntary turnover, social

capital, and organizational performance. Academy

of Management Review, 26, 446–456.

Deutsch, M. (1975). Equity, equality, and need: What

determines which value will be used as the basis

of distributive justice? Journal of Social Issues, 31,

137–149.

Ellwardt, L., Labianca, G., & Wittek, R. (2012). Who

are the objects of positive and negative gossip at

work? A social network perspective on workplace

gossip. Social Networks, 34, 193–205.

Felps, W., Mitchell, T. R., Hekman, D. R., Lee, T. W.,

Holtom, B. C., & Harman, W. S. (2009). Turnover

contagion: How coworkers’ job embeddedness and

job search behaviors infl uence quitting. Academy

of Management Journal, 52, 545–561.

Friedman, R., & Holtom, B. (2002). The effects of net-

work groups on minority employee turnover inten-

tions. Human Resource Management, 41, 405–421.

Gargiulo, M. (1993). Two-step leverage: Managing

constraint in organizational politics. Administrative

Science Quarterly, 38, 1–19.

Gargiulo, M., Ertug, G., & Galunic, C. (2009). The

two faces of control: Network closure and indi-

vidual performance among knowledge workers.


Greenberg, J. (2004). Stress fairness to fare no stress:

Managing workplace stress by promoting or-

ganizational justice. Organizational Dynamics, 33,

352–365.

Griffeth, R., Hom, P., & Gaertner, S. (2000). A meta-

analysis of antecedents and correlates of employee

turnover: Update, moderator tests, and research

implications for the next millennium. Journal of

Management, 26, 463–488.

Grosser, T., Lopez-Kidwell, V., & Labianca, G. (2010).

A social network analysis of positive and negative

gossip in organizational life. Group & Organization

Management, 35, 177–212.

Holtom, B., Mitchell, T., Lee, T., & Eberly, M. (2008).

Turnover and retention research: A glance at the

past, a closer review of the present, and a venture

into the future. Academy of Management Annals,

2, 231–274.

Kaše, R., Paauwe, J., & Zupan, N. (2009). HR practices,

interpersonal relations, and intrafi rm knowledge

transfer in knowledge intensive fi rms: A social net-

work perspective. Human Resource Management,

48, 615–639.

Kilduff, M., & Krackhardt, D. (1994). Bringing the indi-

vidual back in: Structural analysis of the internal

market for reputation in organizations. Academy of

Management Journal, 37, 87–108.

Krackhardt, D. (1990). Assessing the political land-

scape: Structure, cognition, and power in or-

ganizations. Administrative Science Quarterly, 35,

342–369.

Krackhardt, D., & Hanson, J. (1993). Informal networks:

The company behind the chart. Harvard Business

Review, 71(4), 104–111.

Krackhardt, D., & Porter, L. (1986). The snowball effect:

Turnover embedded in communication networks.

Journal of Applied Psychology, 71, 50–55.



Kristof-Brown, A. L., Zimmerman, R. D., & Johnson, E.

C. (2005). Consequences of individuals’ fi t at work:

A meta-analysis of person-job, person-

organization, person-group, and person-supervisor

fi t. Personnel Psychology, 58, 281–342.

Labianca, G., & Brass, D. (2006). Exploring the social

ledger: Negative relationships and negative asym-

metry in social networks in organizations. Academy

of Management Review, 31, 596–614.

Lee, K., & Allen, N. (2002). Organizational citizenship

behavior and workplace deviance: The role of affect

and cognitions. Journal of Applied Psychology, 87,

131–142.

Lee, T. W., & Mitchell, T. R. (1994). An alternative approach:

The unfolding model of voluntary employee turnover.

Academy of Management Review, 12, 51–89.

Lee, T. W., Mitchell, T. R., Sablynski, C. J., Burton, J. P.,

& Holtom, B. C. (2004). The effects of job embed-

dedness on organizational citizenship, job perform-

ance, volitional absences, and voluntary turnover.

Academy of Management Journal, 47, 711–722.

LePine, J. A., Methot, J. R., Crawford, E. R., &

Buckman, B. (2012). A model of positive relation-

ships in teams: The role of instrumental, friend-

ship, and multiplex social network ties. In L. T. Eby

& T. D. Allen (Eds.), Personal relationships: The

effect of supervisory, co-worker, team, customer

and nonwork exchanges on employee attitudes,

behavior, and well-being (pp. 173–194). New York,

NY: Routledge.

Leventhal, G. S. (1976). The distribution of rewards and

resources in groups and organizations. In

L. Berkowitz & W. Walster (Eds.), Advances in

experimental social psychology (Vol. 9, pp. 91–131).

New York, NY: Academic Press.

MacKinnon, D. P., Fairchild, A. J., & Fritz, M. S. (2007).

Mediation analysis. Annual Review of Psychology,

58, 593–614.

March, J. G., & Simon, H. (1958). Organizations. New

York, NY: Wiley.

Meyer, J., Stanley, D., Herscovitch, L., & Topolnytsky,

L. (2002). Affective, continuance, and normative

commitment to the organization: A meta-analysis

of antecedents, correlates, and consequences.

Journal of Vocational Behavior, 61, 20–52.

Meyer, J. P., & Allen, N. J. (1984). Testing the “side-

bet theory” of organizational commitment: Some

methodological considerations. Journal of Applied

Psychology, 69, 372–378.

Mitchell, T., Holtom, B., Lee, T., Sablynski, C., & Erez,

M. (2001). Why people stay: Using job embedded-

ness to predict voluntary turnover. Academy of

Management Journal, 44, 1102–1121.

Mobley, W. (1977). Intermediate linkages in the rela-

tionship between job satisfaction and employee

turnover. Journal of Applied Psychology, 62,

237–240.

Morey, N., & Luthans, F. (1991). The use of dyadic

alliances in informal organization: An ethnographic

study. Human Relations, 44, 597–618.

Mossholder, K., Settoon, R., & Henagan, S. (2005).

A relational perspective on turnover: Examining

structural, attitudinal, and behavioral predictors.


Oldroyd, J. B., & Morris, S. S. (2012). Catching falling

stars: A human resource response to social capi-

tal’s detrimental effect of information overload on

star employees. Academy of Management Review,

37, 396–418.

O’Reilly, C. A. (1980). Individuals and information over-

load in organizations: Is more necessarily better?


O’Reilly, C. A., Caldwell, D. F., & Barnett, W. P. (1989).

Work group demography, social integration, and

turnover. Administrative Science Quarterly, 34,

21–37.

Organ, D. W. (1988). Organizational citizenship behav-

ior: The good soldier syndrome. Lexington, MA:

Lexington Books.

Preacher, K. J., & Hayes, A. F. (2004). SPSS and SAS

procedures for estimating indirect effects in simple

mediation models. Behavior Research Methods,

Instruments, and Computers, 36, 717–731.

Sagie, A., Birati, A., & Tziner, A. (2002). Assessing the

costs of behavioral and psychological withdrawal:

A new model and an empirical illustration. Applied

Psychology: An International Review, 51, 67–89.

Sasovova, Z., Mehra, A., Borgatti, S. P., & Schippers,

M. C. (2010). Network churn: The effects of self-

monitoring personality on brokerage dynamics.


Schneider, B. (1987). The people make the place.

Personnel Psychology, 40, 437–453.

Shah, N. P., Bendersky, C., & Waldstrøm, C. (2011).

Beyond talking shop: Individual performance con-

sequences of multiplex work and social relation-

ships. Unpublished manuscript.

Shaw, J. (2011). Voluntary turnover and organizational

performance: Review, critique, and research agen-

da. Organizational Psychology Review, 1, 187–213.

Shaw, J. D., Duffy, M. K., Johnson, J. L., & Lockhart,

D. E. (2005). Turnover, social capital losses, and

performance. Academy of Management Journal,

48, 594–606.



Shrout, P. E., & Bolger, N. (2002). Mediation in

experimental and nonexperimental studies: New

procedures and recommendations. Psychological

Methods, 7, 422–445.

Sobel, M. E. (1982). Asymptotic confi dence intervals

for indirect effects in structural equation models.

In S. Leinhardt (Ed.), Sociological methodology

(pp. 290–312). San Francisco, CA: Jossey-Bass.

Tett, R., & Meyer, J. (1993). Job satisfaction,

organizational commitment, turnover intention,

and turnover: Path analyses based on meta

analytic fi ndings. Personnel Psychology, 46,

259–293.

Turnipseed, D. L., & Wilson, G. L. (2009). From discre-

tionary to required: The migration of organizational

citizenship behavior. Journal of Leadership &

Organizational Studies, 15, 201–216.

Tushman, M., & Nadler, D. (1978). Information

processing as an integrating concept in organiza-

tional design. Academy of Management Review, 3,

613–624.

Umphress, E. E., Labianca, G., Brass, D. J., Kass, E.,

& Scholten, L. (2003). The role of instrumental and

expressive social ties in employees’ perceptions

of organizational justice. Organization Science, 14,

738–753.

Waller, M. J. (2007). Preferences, behaviors, and strat-

egies for multiple task performance. In

F. Dansereau & F. J. Yammarino (Eds.), Multi-level

issues in organizations and time (pp. 239–248).

Oxford, UK: JAI Press.

SCOTT M. SOLTIS, FILIP AGNEESSENS, ZUZANA ... SOCIAL NETWORK PERSPECTIVE ON TURNOVER INTENTIONS: THE...

Documents

Transcript of SCOTT M. SOLTIS, FILIP AGNEESSENS, ZUZANA ... SOCIAL NETWORK PERSPECTIVE ON TURNOVER INTENTIONS: THE...