A Social Embeddedness Perspective on Turnover Intention: The Role of Informal Networks and Social Identity Evidence From South Korea

Abstract

This study examines turnover intention through a social embeddedness perspective proposing that turnover intention may be a function of the degree to which an organization’s members are attached to one another in terms of relational ties and emotional bonds. Drawing on network theory and social identity theory, it was hypothesized that peripheral positions in informal networks (solidarity ties and instrumental ties) and marginal identity in the workplace may influence higher turnover intention. Sequential mixed methods design was utilized to explore the context-specific bases upon which informal networks and social identities can form and to test the generality of the link between the explored bases and turnover intention against larger samples using Ordered Logistic Model. The results showed that (a) peripherally positioned individuals in informal networks will likely have high turnover intention and (b) individuals with marginal identity in the workplace will likely have high turnover intention. The study results suggest that the social factors accrued from informal networks and social identities deserve enhanced attention in both theorization and personnel management.

Keywords

turnover intention social embeddedness informal networks social identity mixed methods design

Introduction

The topic of voluntary turnover (departure from an organization even though one has the opportunity to stay) has received considerable attention in public management for the following reasons: First, an organization’s high turnover rate might be a sign of ineffective managerial or grievance settlement practices (Cohen, Blake, & Goodman, 2016; cf. Meier & Hicklin, 2008). Second, the recruitment and training of new employees have associated costs and are capable of disrupting shared “tacit knowledge” within an organization (Bertelli, 2007; Moynihan & Landuyt, 2008). Third, high turnover makes long-term human capital planning in an organization difficult (Condrey, 2005). Fourth, the turnover issue has greater significance in management in the context of growing concerns over low birthrates and the resulting retirement wave of the current workforce, which organizations are facing in many countries (Cho & Lewis, 2012; Lewis & Cho, 2011). Previous studies have thus sought to contribute to exploring the determinants of voluntary turnover in their practical importance.

Nevertheless, our knowledge about the determinants of turnover intention is far from complete. This is partly because previous research on turnover in public management has somewhat neglected opportunities to include the social embeddedness perspective, which views organizations as social contexts where members are attached to one another in terms of relational ties and emotional bonds, and these would be the basis upon which individuals choose their course of action, including their intent to leave or stay (Mitchell, Holtom, Lee, Sablynski, & Erez, 2001; Moynihan & Pandey, 2008). Especially with the current trend of increased team-based operations and workplace diversity in management (Choi, 2009), the social embeddedness perspective is likely to gain more importance because, under such contexts, the role of relational ties and emotional bonds in terms of coworker support or emotional affection among colleagues, and thereby in motivating or demotivating employees from attachments to the organization, will likely be more salient (Cole & Bruch, 2006; Oh, Chung, & Labianca, 2004). Nevertheless, while the foci of turnover research have been polarized either into formal factors (e.g., market approaches) or personal factors (e.g., psychological approaches) with regard to behavioral predictors of turnover intention, the fact that humans are essentially social beings whose actions are embedded and guided by their social embeddedness has been afforded insufficient attention thus far, as evidenced by Soltis, Agneessens, Sasovova, and Labianca (2013).

The current study seeks to address this insufficiency, through the lens of the social embeddedness perspective, by investigating how one’s structural position in informal networks and marginal identity in the workplace may influence turnover intention. More specifically, two types of informal networks, “solidarity ties” (e.g., friendship networks) and “instrumental ties” (e.g., advice networks), are considered (Gibbons, 2004), and how peripheral positions in informal networks affect turnover intention is examined using relational network data (cf. Feeley, 2000; Mitchell et al., 2001) drawing on Feeley’s (2000) Erosion Model (EM) and Mitchell et al.’s (2001) Job Embeddedness Model. In addition, this study investigates how marginal and underrepresented individuals have a greater intention to leave when they are likely to experience exclusionary pressures drawing on social identity theory (SIT) and self-categorization theory (Hogg & Terry, 2000). In addition, because informal networks and social identity are characterized by their hidden and implicit nature, there were potential analytical challenges and difficulties (Cross, Nohria, & Parker, 2002). To negotiate such issues, this study utilizes sequential mixed methods design, where the preceding exploratory qualitative phase enables the exploration and building of more context-specific constructs, such as the basis upon which informal networks and marginal identity can form, and a following explanatory quantitative phase validates the generality of the constructs against a larger sample, as suggested by Cresswell and Clark (2007).

Social Embeddedness and Turnover Intention

The “social embeddedness” perspective views organizations as social contexts in which members are attached to one another in terms of “relational ties” and “emotional bonds,” and these ties and bonds are the basis upon which individuals decide courses of action (Mitchell et al., 2001; cf. Polanyi, 1957). As Bourdieu (1984) insightfully suggested, humans are essentially social beings who define their attitudes and actions on their social location in relation to others. Likewise, Granovetter (1985) used the concept of embeddedness to account for how social relations influence human actions.

Looking at turnover through a social embeddedness perspective, it would be a function of the quality of one’s social relationships with others within a workplace. The literature building on this idea tends to link turnover intention to an array of social factors within the workplace such as social cohesion, social integration, social support, and social capital (O’Reilly, Caldwell, & Barnett, 1989). The rationale is that such social factors work as a sort of “social glue” that binds members together, and such social forces can influence ones’ psychological states, such as happiness, satisfaction, and even performance, which may be associated with their intention to leave or stay (Gibbons, 2004, p. 238; Mossholder et al., 2005).

Although the idea of social embeddedness is not entirely novel, the literature, particularly in public management, that empirically links turnover and embeddedness has not received sufficient attention (a notable exception is found in Mitchell et al.’s [2001] Job Embeddedness Model [JEM]). It is partly because most previous research on turnover tended to be polarized to either a “market approach” or a “psychological approach” (Tanova & Holtom, 2008, pp. 1554-1555). As an illustration, the market approach focuses on deterministic external influences as “pulling factors” such as job market conditions, promotion opportunities, job mobility, market information, and managerial programs (e.g., family-friendly policies or New Public Management (NPM) -inspired practices; Bae & Goodman, 2014; Choi, 2009; Kim, 2005; Moynihan & Landuyt, 2008; Selden & Moynihan, 2000). In contrast, the “psychological approach” focuses on personal satisfaction, stress, burnout, personality, and so forth as predictors of “motivation factors” (Steel & Lounsbury, 2009, for a review). These then remind us of Granovetter’s (1985) well-known conceptual distinction between the “over-socialized approach” and “under-socialized approach,” where he suggested that explaining human actions as either obedient to external conditions or personal decisions in isolation might not suffice as they neglect the social ties and bonds that guide human action (Mitchell et al., 2001; Moynihan & Pandey, 2008).

Two theoretical traditions should be examined to operationalize social embeddedness:

1. First, the structuralist tradition that views embeddedness in terms of informal social networks behind formal organizational charts where members exchange various resources such as advice or friendship (Coleman, 1988).

Along with social capital theory, this treats the social networks shared by members as a type of conduit through which social support is exchanged, and those who are marginalized or isolated from such networks are thought to be relatively disadvantaged and thus more likely to intend to leave.

2. Second, social identity tradition that views embeddedness with regard to the degree to which members identify with one another and thereby create emotional bonds with fellow group members (Brown, 2000, for a review).

Social identity theory (SIT) suggests that people in social contexts tend to identify more strongly with others with whom they share more characteristics through social comparison processes, and those who are outsiders or underrepresented in terms of social identity will likely experience exclusionary pressures and are therefore more likely to intend to leave.

Rather than explore “relational ties” (structuralist tradition) and “emotional bonds” (social identity tradition) separately, as is the case with most turnover studies, this study examines how these factors influence turnover intention in a unified model (cf. Mehra, Kilduff, & Brass, 1998).

The Role of Informal Networks

According to the structuralist perspective of embeddedness, informal networks are important sources with which to explain member behavior. Informal networks refer to the patterned configurations of interactions that are emergent, voluntary, and haphazard, not formally designed or prescribed by an organization’s formal structure (Krackhardt & Hanson, 1993; Schalk, Torenvlied, & Allen, 2011). Informal networks are an attractive subject worthy of exploration because they help understand the implicit patterns of human behavior, group processes, and organizational outcomes in workplaces that are not readily visible as they are hidden behind organizational charts but obviously present (Rank, 2008). Informal networks as subjects have great significance as modern organizations are dissolving formal divisional boundaries and promoting ad hoc team operations and inter-unit collaborations as the principles of organizational management (Eberly, Holley, Johnson, & Mitchell, 2011).

Two types of informal networks are commonly identified: (a) “solidarity ties” and (b) “instrumental ties” (Gibbons, 2004; Tichy, Tushman, & Fombrun, 1979). First, “solidarity ties” refer to social relations that are egalitarian, voluntary, trusting, and enduring; a foremost example is friendship networks (Ibarra, 1993). Solidarity ties often begin with an attraction to similar others, and consolidate through shared experiences, frequent interactions, and reciprocity norms (McPherson, Smith-Lovin, & Cook, 2001). Solidarity ties are generally deemed related to Granovetter’s (1985) “strong ties,” which are characterized by strong cohesion and emotional attachment, and thereby serve as “social glue” that reduce the willingness of members to leave their group. Second, “instrumental ties” refer to social relations that help solve problems in cooperation with colleagues by sharing useful resources including information, knowledge, and advice related to work within an organization; a foremost example is “advice networks” (Gibbons, 2004; Sparrowe, Liden, Wayne, & Kraimer, 2001). Previous research has linked the high-quality instrumental ties that individuals develop with their increased opportunities for performance, promotion, and power (Monge & Contractor, 2003). In general, instrumental ties are assumed to give individuals informational advantages and are related to Granovetter’s (1985) “weak ties.”

Network theorists link informal networks and turnover intention based on a “structural equivalence hypothesis,” meaning that the individuals who occupy structurally similar (equivalent) positions within a network may have a potential role in similar behavioral patterns such as turnover (Krackhardt & Porter, 1986). Mossholder et al. (2005) called such hypothesized effects “social embedding effects” (p. 609), and a set of network properties were considered and investigated against turnover. For instance, Feeley (2000) suggested the EM of employee turnover, which argues that individuals whose positions are structurally peripheral (or marginal) with lower centrality in an informal network tend to “fall off” more easily from an organization. Consistently, Mossholder et al. (2005) found that members’ centrality and turnover intention were negatively associated in a health care medical center setting. In addition, Krackhardt and Porter (1986) found a similar pattern in a fast-food restaurant setting in an earlier study. Regarding such social networks’ influence on turnover, there are two theoretical explanations depending on which type of informal network is focused upon.

First, those who point to “solidarities ties” (e.g., friendship networks) understand the benefits accrued via informal networks from affective and emotional relief. This is related to the concept of “workplace social capital,” which refers to the degree to which group members are cohesive in terms of their reciprocal relationships (Coleman, 1988; Westphal & Milton, 2000). It points to the importance of coworker support or emotional affection as intangible assets that are conducive to solidarity benefits such as social norms and happiness, and thereby reduces turnover intention (Coleman, 1988). In support of this argument, Moynihan and Pandey (2008) and Bertelli (2007) found that when members experience solidarity in an organization, this reduces turnover intention. Taken together, the following hypothesis is proposed based on these findings.

Hypothesis 1 (H1): Individuals with more peripheral positions in solidarity ties (friendship networks) within organizations are likely to have higher turnover intention.

Second, according to the job demands–resources (JD-R) model, individual performance depends on both the level of job demand and the resources available to meet the demand, and informal ties reportedly serve as a type of “resource channel” (Demerouti, Bakker, Nachreiner, & Schaufeli, 2001). It points to the importance of instrumental ties in terms of their advantages in finding solutions to task-related problems. Enhanced problem-solving capacities are likely conducive to the sense of accomplishment and satisfaction, which will likely lead to a reduction in turnover intention (Burt, 2000). Then, those individuals who are positioned peripherally in instrumental ties (or advice networks) may be moved away from the task-related status quo and are therefore disadvantaged (Gibbons, 2004). The following hypothesis is proposed based on these arguments (see Oh et al., 2004; Soltis et al., 2013, for a review).

Hypothesis 2 (H2): Individuals with more peripheral positions in instrumental ties (advice networks) within organizations are likely to have a higher turnover intention.

The Role of Social Identity

SIT suggests that in a social context, individuals tend to identify more strongly with others with whom they share similar characteristics (e.g., gender, age, preference, role, and status, etc.) through social comparison processes (Hogg & Terry, 2000). Such social comparison and identification processes permit the formation of boundaries between the in-group and relevant out-groups, in a process called self-categorization (Brown, 2000; Turner, Hogg, Oakes, Reicher, & Wetherell, 1987). Social identity theorists posit that human behaviors are guided by the self-images that have been constructed through such processes, and human behaviors are embedded in social contexts.

SIT may have provided a good explanation of the theoretical mechanisms for how certain individuals are marginalized or discriminated against within social settings such as the workplace within organizations (Mehra et al., 1998). For instance, according to homophily hypothesis and the similarity-attraction hypothesis, people tend to favor similar others in terms of social categories and identities, because they reduce psychological discomfort and ease communication (Monge & Contractor, 2003). The reasoning is represented well in the saying “Birds of a feather flock together” (cf. McPherson et al., 2001), and the mechanisms explain why in-group bias (or favoritism) occurs (Tsui, Egan, & O’Reilly, 1992; Westphal & Stern, 2007).

Social identity theorists have empirically focused on the link between in-group bias and turnover intention in organizational contexts (Brown, 2000; Cole & Bruch, 2006). Distinctiveness theory is particularly noteworthy, as it provides a parsimonious answer about how relative rarity in terms of social identity in the workplace can engender unfair treatment, alienation, or discrimination (Mehra et al., 1998). According to the distinctive theory, those who have relatively rare characteristics are likely more salient and visible due to their distinctive nature, rendering them more susceptible to being defined as a marginal out-group (Cole & Bruch, 2006). In a similar vein, social-psychologists call such minority effects “black sheep effects” (Marques, Yzerbyt, & Leyens, 1988). Furthermore, Blumer’s (1958) Group Threat Theory proposes that marginal individuals in terms of gender, race, ethnicity, and so on are susceptible to faulty stereotypical images or prejudices and to social threats such as hostility exerted by a dominant identity (Dixon, 2006).

There is substantial research arguing that when there are marginal or underrepresented individuals in a workplace, they are more likely to experience such exclusionary pressures and thus have higher intention to leave (O’Reilly et al., 1989; Steel & Lounsbury, 2009; Tsui et al., 1992). Individuals within an organization are also likely to create in-groups with others similar to themselves, creating a psychological barrier against the out-group (McPherson et al., 2001). When such an in-group/out-group distinction forms in the workplace, marginal members who are most different will likely have lower satisfaction and loyalty toward either coworkers or the organization, which may lead to increased turnover intention; accordingly, the following hypothesis is proposed.

Hypothesis 3 (H3): Individuals with higher marginal identity within the workplace are likely to have higher turnover intention.

Method

Research Context

This study was conducted with research-type public agencies in Korea. This organization type was chosen based on the following reasons. First, this type of agency hires half of their employees on a contract basis, unlike other public employees, and consequently has lower job security, which makes the agencies more vulnerable to employee turnover problems. In contrast, employees in other public sectors in Korea generally have a high level of job security and even guaranteed permanent employment; as a result, they have a very low level of work mobility (Im, Campbell, & Cha, 2013). In this social context, research-type public agencies are considered an exception to the rule.¹ Second, the other half of the employees in these agencies who hold noncontractual positions are mostly researchers, and they have more job alternatives and a higher level of job mobility compared with employees in other public agencies. For example, researchers generally hold an advanced degree in a specialized discipline and have relatively many job alternatives, including educational institutions, research firms, and private enterprises. Third, research-type public agencies conduct research requiring a high level of expertise and advanced technology and have a large proportion of long-term projects that are conducted by research teams (Van den Bulte & Moenaert, 1998). As a result, the issue of cost due to disruption in work and retraining, as a result of employee turnover, is more pronounced. These factors make research-type public agencies good candidates for turnover research.

Meanwhile, the research content in Korea has a distinctive advantage in social embeddedness research. Korea is a country with a collectivistic culture (others include Taiwan, China, and Japan), in which strong social ties among members are valued and members find meaning and motivation for social behaviors (Hofstede, 2001).² These countries are good candidates for conducting research on relations among organization members because of their strong emphasis on networks and cohesion (Nisbett, 2010; Triandis, 1995). In fact, East Asian countries, including Korea, have distinct concepts of cohesive networks based on mutual reciprocity, such as inmak (Korea), guanxi (China), and kankei (Japan; Guthrie, 1998; Hitt, Lee, & Yucel, 2002). All of these aspects of Korea offer a useful setting for research on social embeddedness effects on turnover intention.

Research Procedure and Data

I conducted the study in two sequential phases. (a) The first phase was in-depth exploratory interviews with public employees who had worked or were working in research-type agencies. The goal of the phase was to understand and identify the categorical bases upon which informal networks and self-categorization are defined within the agencies. The bases are context-specific according to individual organizations and cultures (Siegel, 2007). For instance, college alumni may be a salient basis of informal network formation in some organizations, while religious affiliation may be in others. Also, in female-dominant organizations, male employees are defined as a marginal actor and vice versa in male-dominant organizations. This suggests that, without a context-specific understanding of the categorical bases in the agencies, it is impossible to construct accurate survey instruments for collecting relational data for network analysis (Gibbons, 2004; Uzzi, 1997). This makes the first phase necessary. (b) The second phase included construction of a survey based on the preceding interviews and testing the proposed hypotheses using a larger sample to validate the generality of the theory (Cameron, 2009).

In other words, the primary design of this study is the two phases sequentially and organically linked together: the preceding interviews for conceptual exploration and the following surveys for generality validation (Cresswell & Clark, 2007). This resembles the sequential mixed methods design proposed by Cameron (2009). The following section provides details of each phase.

Phase 1: Preceding interviews and findings

The in-depth exploratory interviews were conducted in summer 2015 with 19 public employees who had worked or were working in research-type agencies. The participants were selected using purposive sampling based on the criteria of a researcher or a manager with seniority and their experience in respective research areas, or an employee with a major responsibility in human resource management (Choi & Park, 2014).³ Initially, through the purposive sampling, 10 potential participants were identified and contacted, six of them who consented to participate were interviewed, and nine more participants were recruited using the snowballing procedure based on the initially interviewed participants’ referrals, amounting to a total number of 19 interviews.

The questions in the interviews fell into three categories: (a) “the basis of solidarity ties (e.g., friendship networks) in the organization,” (b) “the basis of task-related instrumental ties (e.g., advice networks) in the organization,” and (c) “the basis of marginal identity within the workplace.” The content of the interviews was transcribed into textual data using word processing software and imported into ATLAS.ti v7.5, qualitative data analysis (QDA) software for analytic efficiency. The textual data were then analyzed according to a “thematic analysis” derived from Braun and Clarke’s (2006) QDA technique. The thematic analysis is a qualitative coding method that identifies meaning segments, such as words, sentences, and paragraphs, that correspond to information that the investigator wants to know and reports themes based on categorization of the meaning segments (Braun & Clarke, 2006; Saldaña, 2012).

Based on the thematic analysis, the following bases of informal networks and marginal identity were identified (Appendix 2). First, the bases of relationship building in terms of solidarity ties were college alumni, region-based social circles, and hobby communities. The respondents identified college alumni associations and region-based social circles as the most influential bases of informal network building within organizations, and most respondents were found to strengthen solidarity and cohesion through regular informal meetings. Here, region refers to birthplace, and hobby communities refer to self-organized informal groups in organizations based on shared hobbies or interests such as sports, travel, music, art, and foreign languages. Second, the bases of relationship building in terms of task-related instrumental ties were cohorts and research collaborators. Cohorts are those who join an organization in the same year and were found to frequently exchange resources needed for work, such as information and advice. Research collaborators are a group of coworkers who form voluntarily to conduct joint research projects outside the categories of formal assignments. They are highly similar in the areas of interest and methodologies they use and were also found to actively exchange resources needed for formal assignments. Third, the categories that are defined in the workplace were found to be gender, functional background, and education level. While gender is a traditional category in demography studies (e.g., Moynihan & Landuyt, 2008), functional background and education level are relatively nascent categories which are explored in this study. In contrast, other traditional categories, including race, handicap, religion, and marital status, were not found to be major issues in the organizations.

Phase 2: Following survey and instruments

The second phase involved testing the proposed hypotheses and validating the theory, for which a survey was conducted with a larger sample. Of the 420 comprising the sample (four organizations), 240 responded, resulting in a response rate of 57.0% (129 male; 111 female). The survey was constructed with the following two sets of questionnaires. The first set consisted of self-reporting instruments on respondents’ attributes, including turnover intention, job satisfaction, workload, gender, functional background, educational level, and so on. The second set consisted of relationality instruments in the form of “who-to-whom” questions to analyze the position each individual occupies in informal networks. For instance, the question on college alumni association as one of the categories identified based on the preceding interviews is “Place a check mark next to the individuals you interact with at the College Alumni Association on a regular basis.” A question in a similar format was asked for each identified category (Feeley, 2000). Relational data obtained from each individual were aggregated into network data, which are defined as below.

Network data are in square data matrices (the same number of rows and columns), and the rows and columns of the matrices are assigned the names of respondents (nodes), and the cells of the matrices contain 1 or 0 entries that represent whether there is a relation between all possible pairs of nodes. (Monge & Contractor, 2003, p. 36)

The network data were then imported into UCINET software for the calculation of ego-centric centrality per respondent (Borgatti, Everett, & Freeman, 2002).

Dependent Measure

Turnover intention as a dependent variable was measured in two dimensions—long-term and short-term—based on Moynihan and Pandey’s (2008) study. Short-term turnover intention was measured with the question “How often do you look for job opportunities outside this organization?” (1 = never, 5 = constantly), and long-term turnover intention was measured with “I would be happy to spend the rest of my career with this organization” (7 = strongly disagree, 1 = strongly agree). Most previous research tended to use short-term turnover intention only as a standard measure with a single statement of leaving intention (see Steel & Lounsbury, 2009, for a review). However, according to Meyer and Allen (1991), when measuring attitudinal variables (e.g., commitment), it is clearer to make a conceptual distinction between short-term and long-term intention because short-term intention would largely be related to the motivation of status enhancement, whereas long-term intention would involve normative propensity or moral obligation. Accepting these claims, short-term turnover intention can be then seen as an expression of actionable job dissatisfaction, while long-term turnover intention can be related to a sense of loyalty to the organization (cf. Moynihan & Pandey, 2008). Another issue pertaining to turnover intention as a dependent variable is whether turnover intention is a reasonable proxy for actual turnover (Cohen et al., 2016; Dalton, Johnson, & Daily, 1999). Regarding this issue, Cho and Lewis (2012, p. 19) view turnover intention as a “reasonable proxy for actual turnover” (Campbell & Im, 2016; Cho & Lewis, 2012). The author agrees with this view and furthermore concludes that using turnover intention as a proxy was suitable in the present study due to the advantage that the proxy allows the use of cross-sectional data.

Independent Measures

Network periphery

To assess the degree of peripheral positions an individual occupies in informal networks, the degree of normalized centrality, ranging from 0 to 1, was calculated for each individual based on the collected network data and then subtracted from 1. The degree centrality of a node refers to the number of ties attached to the node and is a proper indicator for measuring the strength of the ties an individual has, as it indicates the number of contacts of the focal actor (Bonacich, 1987; Feeley, 2000).⁴ Meanwhile, the degree centrality can be either in-degree centrality or out-degree centrality depending on the direction of a relation, and this study used the degree centrality without differentiation of its direction as an independent measure. The decision was made following Feeley’s (2000) recommendation that the directionality of relational data must be removed from the measurements for network data collected with less than a 100% response rate. The formula for the focal actor (ego)’s degree periphery is as follows:

1 - \frac{1}{n - 1} • \sum_{i, j = 1}^{n} X_{i} \cdot X_{j} .

To illustrate, the measure is calculated by the degree of focal actor’s normalized centrality and then subtracted from 1. Here, the centrality measure is “the value obtained by dividing the total number of contacts (Xij) a focal actor (Xi) has by the number of members of the network excluding oneself” and takes a value of between 0 and 1. Thus, if the value is subtracted from 1, then 1 indicates a sole isolate and 0 indicates complete connectivity (Feeley, 2000).

Marginal identity

To measure the relative degree of marginality (or rarity) of an individual in the workplace, this study used the relational score proposed by O’Reilly et al. (1989) and Tsui et al. (1992), and the score was calculated for each identified gender, functional background, and education level (O’Reilly et al., 1989; Tsui et al., 1992). Although marginal identity can be simply measured as a dummy variable, it was determined to measure the variable as a degree of psychological experience as more valid in the present study, considering one’s psychological experience of being a marginal individual in the workplace is likely to have an attribute of the degree (Westphal & Milton, 2000). For instance, for a woman working in a workplace with one female and three male employees and a woman working in a workplace with one female and 20 male employees, there is likely to be a difference in the strength of their psychological experience as a minority. Moreover, individual characteristics that determine rarity vary across contexts. For instance, although women have been more commonly defined as a minority gender than men, assuming a female-dominant workplace, it is then reversely more reasonable to define men as a minority gender within the workplace. The relational score was designed to incorporate these aspects of the degree of psychological experience and the relativity in determining the status. The variable is expressed in the following formula:

{[\frac{1}{n} \sum_{j = 1}^{n} {(X_{i} - X_{j})}^{2}]}^{\frac{1}{2}} .

To illustrate, this relational score is calculated by “the square root of the summed squared differences between an individual Xi’s value on a specific category (e.g., gender) and the value on the same category for every other individual Xj’s in the sample, divided by the total number (n) of respondents in the work unit” (O’Reilly et al., 1989, p. 25; Tsui et al., 1992, p. 562). The relational score takes a value of between 0 and 1, but never reaches 1.00. An individual who scores 0.999 indicates sole minority in an enormously large group.⁵

Controls

Some other variables are examined as controls due to their own practical importance. First, previous research has argued that family-friendly policies are conducive to reduced turnover intention and absenteeism (Bae & Goodman, 2014; Lee & Hong, 2011). As with the current trend that values a healthy work–life balance (WLB), public agencies have implemented an array of family-friendly policies including child care subsidies, paid leave for family-care, teleworking or alternative work schedules, and so on (Huang, Lawler, & Lei, 2007). Thus, the degree of such policy supports felt by employees were included as a control and measured with a 5-point Likert-type scale (Cohen et al., 2016; Lee & Hong, 2011). Second, previous research has consistently argued that job satisfaction in terms of payment and workload is a major determinant of turnover intention (Ertas, 2015; Reid & Nygren, 1988; Tsui et al., 1992). I therefore controlled for both pay satisfaction and workload satisfaction using a 5-point Likert-type scale to measure to what extent employees felt satisfied with their remuneration and workload. Third, other researchers have also suggested that a time variable such as years in position is associated with turnover intention (Cho & Lewis, 2012; Moynihan & Landuyt, 2008). According to this argument, older employees are less likely to intend to quit due to increased family obligations, position settlement, pension plans, and do on. Moynihan and Landuyt (2008) call this the life cycle stability hypothesis. It was controlled for and measured by asking employees for their seniority in years. Fourth, the value fit between a person and organization (P-O value fit) was included as a control, because there is substantial empirical support that compatibility and matches between people and organizations reduce turnover intention (Mossholder et al., 2005; Moynihan & Pandey, 2008). The P-O value fit was measured on a 5-point Likert-type scale. (Refer to the appendix 1 for a detailed description of variables and survey statements.)

Statistical Analysis

Ordered Logistic (or Ordered Logit) Regression was employed because of the ordinal nature of the dependent variable (Long & Freese, 2001). In general, when a dependent variable (y) is a qualitative variable with ordinal categories, the distances among categories may not be equal. In addition, if ordinary least square (OLS) regression based on the assumption of the dependent variable with consecutive integers or an interval scale is performed on the variable with unequal distances among categories, the probability distribution of the dependent variable (y) may lead to incorrect conclusions due to violation of the assumption of OLS (Long & Freese, 2001).⁶ The dependent variable in this study, turnover intention, has the aforementioned categorical nature, as it is based on respondents’ self-reports based on their qualitative judgment for the sake of analysis convenience. If there is a significant difference in the relationship between independent variables and the logits for all of the logits, the parallel regression assumption for ordinal regression is not satisfied (Moynihan & Pandey, 2008). Thus, to assess model adequacy, a likelihood ratio (LR) test based on chi-square distribution was performed (null hypothesis: no significant difference; Cohen, Cohen, West, & Aiken, 2013). The test results showed that for short-term turnover intention, χ² = 41.931 with p value = .711, and for long-term turnover intention, χ² = 43.495 with p value = .813. This means that the results fail to reject the null hypothesis, suggesting that the lines or planes are parallel and the model meets the assumption of ordinal regression.

Results

Table 1 provides the means, standard deviations, minimum values, and maximum values of the variables. The results of the ordered logistic analysis of turnover intention as a function of social embeddedness factors are shown in Table 2, where the coefficients (B), standard deviation (SD), and odds ratio (EXP(B)) for each independent variable are suggested. In Table 2, there are two models based on dependent variables: short-term turnover intention (Model 1) and long-term turnover intention (Model 2).

Table 1.

Descriptive Statistics.

Variables	Minimum	Maximum	M	SD
Short-term turnover intention	1	5	2.89	1.06
Long-term turnover intention	1	7	3.12	1.28
Periphery in college alumni networks	0	1	0.58	0.16
Periphery in region-based social circles	0	1	0.61	0.14
Periphery in hobby communities	0	1	0.48	0.07
Periphery in cohorts workers	0	1	0.55	0.21
Periphery in joint research collaborators	0	1	0.36	0.17
Minority in gender	0	0.78	0.62	0.09
Minority in functional background	0	0.55	0.27	0.11
Minority in education level	0	0.64	0.31	0.13
Family-friendly policies	2	10	7.18	1.79
Pay satisfaction	4	20	11.09	2.34
Workload satisfaction	2	10	4.09	2.41
Years in position	1	34	12.02	6.83
P-O value fit	4	20	12.21	2.79

Table 2.

Ordered Logit Analysis.

	Short-term turnover intention (Model 1)			Long-term turnover intention (Model 2)
	B	SE	EXP(B)	B	SE	EXP(B)
Periphery in college alumni associations	1.0651 ***	.7483	1.9385	1.0138 ***	.7141	2.0001
Periphery in region-based social circles	1.0982 ***	.5431	2.8991	1.3187 ***	.4181	2.9999
Periphery in hobby communities	.1901 *	.5442	1.2663	.9981 **	.5333	1.3391
Periphery in cohort workers	.3813 **	.4601	1.2842	.7987 *	.4878	1.0111
Periphery in joint research collaborators	.0156	.0180	1.1971	.0206	.0161	1.1310
Marginal identity in gender	1.2647 ***	.7648	2.9990	1.1102 ***	.8143	1.2714
Marginal identity in functional background	.3771*	.4171	1.1265	.9881 **	.5489	1.7131
Marginal identity in education level	.9998*	.6431	1.0157	.8814 *	.6124	1.0011
Family-friendly policies	−0.3547 *	.5614	.6015	−0.4101	.5583	.4987
Pay satisfaction	−0.2999 ***	.4103	.7811	−0.2955 **	.3417	.7660
Workload satisfaction	−0.0480	.0600	.9812	−0.0359 *	.0512	.8810
Years in position	−0.6691 *	.2999	.3133	−0.3104 **	.5131	.5999
P-O value fit	−0.2841 *	.3319	.8101	−0.2931 *	.3481	.7988
Estimated cut points:
τ1	−6.0189	1.0021		−10.490	1.135
τ2	−4.5100	.6859		−8.413	1.004
τ3	−2.1213	.5421		−5.344	.713
τ4	1.0020	.3613		−3.170	.450
τ5				−0.181	.331
τ6				1.358	.043
Number of observations (N)	240			240
Likelihood-ratio χ² (13)	54.20***			55.33***
McFadden adjusted R²	0.223			0.208

Note. Significant Level, *p ≤ .1. **p ≤ .05. ***p ≤ .01.

H1 and H2 predicted a positive association between one’s peripheral positions (1 − ego’s degree centrality) within informal networks and turnover intention. The type of informal network H1 refers to is solidarity ties (e.g., friendship networks), while H2 refers to instrumental ties (e.g., advice networks). In general, the results support the hypotheses.

In particular, in Model 1 where the dependent variable is the short-term turnover intention, H1 predicted that the peripheral positions in solidarity ties would be positively related to turnover intention. Of the bases of such solidarity ties, college alumni associations were significantly related to turnover intention (B = 1.0651, p ≤ .01), region-based social circles were significantly linked to turnover intention (B = 1.0982, p ≤ .01), and hobby communities were significantly related to turnover intention (B = 0.1901, p ≤ .1). These findings are consistent with Siegel’s (2007) argument that region (birthplace) and alumni are the primary bases of informal networks in South Korea, given that the magnitude of these two variables are relatively higher than those of other variables.

In addition, H2 predicted that one’s peripheral positions in instrumental ties would be positively linked to turnover intention, and cohort workers and joint research collaborators were examined. The results showed that one’s peripheral position relative to cohort workers was significantly associated with turnover intention (B = 0.3813, p ≤ .05), while joint research collaborators were statistically insignificant. It is intriguing that among all informal network variables including solidarity ties, joint research collaborators was the only type of network that was found to have a statistically insignificant relationship with turnover intention. The distinct characteristic of the informal network of joint research collaborators is that its formation and maintenance appear not necessarily based on the presence of face-to-face interactions unlike other informal networks. According to the respondents, most communications tend to be carried out online.

Finally, H3 predicted a positive association between marginal identities within the workplace and turnover intention. Of the social categories, gender turned out to have a positive impact and was statistically significant (B = 1.2647, p ≤ .01), and other variables were all statistically significant, but had relatively smaller magnitudes, such as functional background (B = 0.3771, p ≤ .1) and education level (B = 0.9998, p ≤ .1). The results imply that, in South Korea, gender is the most salient factor that defines marginal identity and marginal gender might be more susceptible to exclusionary pressures than other identity factors.

Model 2 is parallel with Model 1 as the same set of independent variables is included; the only difference lies in the fact that Model 2 uses long-term turnover intention, whereas Model 1 uses short-term turnover intention. In particular, H1 predicted peripheral positions in solidarity ties would be positively related to turnover intention. The results showed that college alumni associations were significantly related to turnover intention (B = 1.0138, p ≤ .01), region-based social circles were significantly linked to turnover intention (B = 1.3187, p ≤ .01), and hobby communities were significantly related to turnover intention (B = 0.9981, p ≤ .05). Compared with Model 1, the magnitudes of hobby communities and region-based social circles in Model 2 are larger.

In addition, regarding H2, which predicted one’s peripheral positions in instrumental ties would be positively linked to turnover intention, cohort workers and joint research collaborators were examined. The results showed that one’s peripheral position relative to cohort workers was significantly associated with turnover intention (B = 0.7787, p ≤ .1), while joint research collaborators were statistically insignificant. These results are consistent with those of Model 1.

Finally, H3 predicted a positive association between marginal identities within the workplace and turnover intention. Of the social categories, gender is positively related to turnover intention and statistically significant (B = 1.1102, p ≤ .01). Functional background is positively linked to turnover intention and statistically significant (B = 0.9881, p ≤ .05), and education level (B = 0.8814, p ≤ .1). The findings regarding H3 are generally consistent with those in Model 1.

Overall, the results indicate that of the independent variables that are examined in the models, college alumni associations, region-based social circles, and marginal identity regarding gender are the most salient social factors.

Discussion

This study examines turnover intention through the theoretical lens of a social embeddedness perspective that proposes that turnover intention may be a function of the degree to which an organization’s members are attached to one another in terms of relational and emotional ties (Mitchell et al., 2001; Mossholder et al., 2005). Overall, the findings provide evidence that both one’s structural positions in informal networks and social identities regarding relative marginality in the workplace influence turnover intention. More specifically, the findings indicate that, particularly in South Korean contexts, college alumni associations, region-based social circles, and marginal identity are the most salient categories by which individuals’ attitudes and behaviors are guided. These findings are largely consistent with previous empirical research that explored informal social forces in South Korea—the most notable example is found in the research of Siegel (2007). In addition, the findings support theoretical insights, such as Feeley’s (2000) EM, Mitchell et al.’s (2001) Job Embeddedness Model, Demerouti et al.’s (2001) JD-R Model, and Marques et al.’s (1988) Black Sheep Effect. Thus, not only has this study sought to integrate seemingly fragmented theoretical pieces, but it has also pointed to the practical importance of understanding social aspects within the workplace as a social setting. Our general perspective on management tends to involve dealing with formal arrangements through the well-structured organizational chart of business processes (Cross et al., 2002). However, as the findings of this study showed, even though they are not easily observable or readily visible, social aspects, such as relational and emotional bonds, embedded in the workplace have substantive effects on members’ attitudes. The study then offers several practical and theoretical suggestions for the effective management of employee turnover:

First, managers must consider informal and social aspects, such as informal networks and the social composition of the organization, as important issues (cf. Barnard, 1938). Admittedly, in the observed results in this study, the effects of college alumni networks and region-based social circles on turnover intention were greater than those of pay satisfaction or workload. This suggests that turnover intention derives from psychosocial motives as much as financial motives. These psychosocial motives often result from informal relationships and the compositions of organization members. Therefore, to be more effective, managers need to be sensitive to informal relational aspects within organizations in their managerial approaches (Moynihan & Pandey, 2008).

Second, the finding that informal networks are likely to reduce turnover intention suggests that informal networks within the workplace have the effect of binding members together (Mossholder et al., 2005). This is effectively demonstrated by the finding that individuals who are central to hobby communities have low turnover intention, which suggests that informal small groups within the workplace have certain positive affective effects. This implies that human resource management could be strengthened in practice and in theory by better reflecting the informal networks within the organization, as suggested by Cross et al. (2002) and Soltis et al. (2013). In addition, this study proposes the importance of members’ identification processes to their attitude and behavior. However, it is difficult to follow a fixed criterion when “On what bases informal networks are defined” or “On what criteria marginal identities are defined” are socially constructed according to context-specific conditions (cf. Meyer, 2006). In this regard, this study suggested a mixed-method research design with a qualitative phase for exploring and identifying the key factors latent in the workplace based on interactional processes (i.e., in-depth interviews) with members. I believe that this research strategy enables more context-specific theoretical constructs to be built. In addition, this study serves as a precedent in terms of research design when dealing with similar subtle topics about the informal aspects of organizations.

Despite its theoretical contributions, this study also has several limitations. First, this study was conducted with employees in specific organizations (research-type public agencies) in South Korea, and the study results need to be interpreted with consideration of this special context. Therefore, practitioners and theorists in other countries are advised to take the findings of this study to verify their generalizability rather than as they are (Cameron, 2009). Second, Korea is a highly racially homogeneous country. Consequently, demographic categories, such as race and ethnicity, were not given much emphasis in this study. This can be somewhat different from situations in countries with highly multicultural contexts and where the issues of racial and ethnic minorities in the workplace are thus of great significance. Therefore, this aspect needs to be given attention in follow-up studies that compare different racial and ethnic minorities.

Footnotes

Appendix 1

Variables and Survey Questions.

Variables		Survey questions
Dependent variable
Turnover intention	Short-term turnover intention	“How often do you look for job opportunities outside this organization?” [5 = constantly, 4 = very often, 3 = sometimes, 2 = not very often, 1 = never] (Moynihan & Pandey, 2008)
Turnover intention	Long-term turnover intention	“I would be happy to spend the rest of my career with this organization.” (reversed) [1 = strongly disagree, 7 strongly agree] (Moynihan & Pandey, 2008)
Independent variable
Informal networks	College alumni networks	“On the following pages, you will find a list of all the employees of the organization. Place a check mark next to the individuals you have an intimate relationship with at ‘the affiliations listed below.’ If you do not speak to the person on a regular basis, please do not place a check mark next to his or her name.” (Feeley, 2000; Mossholder et al., 2005) 1) Region (birthplace)-based social circles: ______, ______, ______ 2) College alumni networks: ______, ______, ______ 3) Hobby community: ______, ______, ______ 4) Cohort workers: ______, ______, ______ 5) Joint research collaborators: ______, ______, _____
	Region-based social circles
	Hobby community
	Cohort workers
	Joint research collaborators
Minority status	Gender	Gender of respondents 1) Female, 2) Male
	Functional background	Functional background of respondents 1) Management, 2) Humanity, 3) Philosophy, 4) Engineering, 5) Arts, 6) Natural Science, 7) Law
	Education level	Education level of respondents 1) High-school graduate, 2) College graduate, 3) Master’s degree, 4) PhD degree
Control variables
Family-friendly policies		Summative indexes of responses to the following statements (Cronbach’s α = .87; Moynihan & Landuyt, 2008) 1) “When possible, alternative work schedule (flextime, compressed work weeks, job sharing, telecommuting, etc.) are offered to employees.” [5 = strongly agree, 1 = strongly disagree]
		2) “My organization supports my need to balance work and other life issues” [5 = strongly agree, 1 = strongly disagree]
Pay satisfaction		Summative indexes of responses to the following statements (Cronbach’s α = .67)[5 = strongly agree, 1 = strongly disagree]1) “Considering everything, how satisfied are you with the pay you receive for your job?” (Ertas, 2015; Tsui, Egan, & O’Reilly, 1992)2) “If your team performs well how likely is it that you will receive a cash award or pay increase?” (Cho & Lewis, 2012)3) “If I perform well, it is likely I will receive a cash award or pay increase.”4) “My organization takes steps to ensure that employees are appropriately paid and rewarded.”
Workload satisfaction		Summative indexes of responses to the following statements. (Cronbach’s α = .71)1) “How often do you feel burnout with your work?” [5 = very frequently, 1 = never] (Reid & Nygren, 1988)2) “Your workload is fair amount that you are capable of taking care of with ease” [5 = strongly disagree, 1 = strongly agree] (Robert & Hockey, 1997)
Years in position		“How many years have you been in your current position?” (Moynihan & Pandey, 2008)
Value fit		Summative indexes of responses to the following statements (Cronbach’s α = 0.85) [5 = strongly agree, 1 = strongly disagree] (Moynihan & Pandey, 2008; Tsui et al., 1992)1) “I find that my values and the organization’s values are similar.”2) “If the values of this organization were different, I would not be attached to this organization.”3) “What this organization stands for is important to me”4) “This reason I prefer this organization to others is because of what it stands for, its values.”

Appendix 2

Summary of Interview Findings on Features of Informal Networks and Marginality Identity.

Bases of Informal Network and Marginal Identity(themes)	Description(feature)	Counts
Solidarity Networks
College Alumni Networks	Informal social groups consisting of respective college alumni, which generally operate respective SNS groups and hold regular organized meetings based on member fees.	27
Region-based Social Circles	Informal social groups consisting of those from the same birthplaces, which are close to relatively unorganized friendship ties, and members meet through irregular meetings.	38
Hobby Communities	Social groups organized around common interests, such as music, art, reading, travel, and exercise, which can be either formally organized by the organization or voluntarily organized by individuals.	14
Instrumental Networks
Cohort Workers	Informal social groups consisting of cohorts who have joined the organization in the same year, which generally operate respective SNS groups to share workplace-related news and information.	11
Joint Research Collaborators	Informal networks consisting of those who have participated in respective joint research together, which do not always require face-to-face interaction.	13
Marginal Identities
Gender	Classification of men and women; a work unit can be either female- or male-dominant depending on the male–female ratio.	31
Functional Background	The area of specialization in research and work responsibility, such as biology, chemistry, international trade, accounting, and law.	15
Education Level	Whether a degree in higher education or credentials in elite education are possessed. A work unit generally has a mix of members of various education levels.	22

Note. Numbers in cells represent frequency of responses by interviewees aggregated across person cases.

Declaration of Conflicting Interests

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

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The research was partially supported by the National Research Foundation of Korea Grant funded by the Korean Government (NRF-2016S1A3A2925463).

Notes

Author Biography

Hyang Won Kwon is an adjunct professor and senior researcher in the Graduate School of Governance at the Sungkyunkwan University (SKKU) in Seoul, South Korea. His current research focuses on organization theory, public philosophy, qualitative/quantitative research methods, and government reforms.

References

Bae

K. B.

Goodman

(2014). The influence of family-friendly policies on turnover and performance in South Korea. Public Personnel Management, 43, 520-542.

Barnard

C. I.

(1938). The functions of the executive. Cambridge, MA: Harvard University Press.

Bertelli

A. M.

(2007). Determinants of bureaucratic turnover intention: Evidence from the department of the treasury. Journal of Public Administration Research and Theory, 17, 235-258.

Blumer

(1958). Race prejudice as a sense of group position. The Pacific Sociological Review, 1, 3-7.

Bonacich

(1987). Power and centrality: A family of measures. American Journal of Sociology, 92, 1170-1182.

Borgatti

S. P.

Everett

M. G.

Freeman

L. C.

(2002). UCINET for Windows: Software for social network analysis. Harvard, MA: Analytic Technologies.

Bourdieu

(1984). Distinction: A social critique of the judgement of taste. Cambridge, MA: Harvard University Press.

Braun

Clarke

(2006). Using thematic analysis in psychology. Qualitative Research in Psychology, 3, 77-101.

Brown

(2000). Social identity theory: Past achievements, current problems and future challenges. European Journal of Social Psychology, 30, 745-778.

10.

Burt

R. S.

(2000). The network structure of social capital. Research in Organizational Behavior, 22, 345-423.

11.

Cameron

(2009). A sequential mixed model research design: Design, analytical and display issues. International Journal of Multiple Research Approaches, 3, 140-152.

12.

Campbell

J. W.

(2016). PSM and turnover intention in public organizations: Does change-oriented organizational citizenship behavior play a role? Review of Public Personnel Administration, 36, 323-346.

13.

Cho

Y. J.

Lewis

G. B.

(2012). Turnover intention and turnover behavior: Implications for retaining federal employees. Review of Public Personnel Administration, 32, 4-23.

14.

Choi

(2009). Diversity in the US Federal Government: Diversity management and employee turnover in federal agencies. Journal of Public Administration Research and Theory, 19, 603-630.

15.

Choi

Park

C.-O.

(2014). Glass ceiling in Korean Civil Service analyzing barriers to women’s career advancement in the Korean government. Public Personnel Management, 43, 118-139.

16.

Cohen

Blake

R. S.

Goodman

(2016). Does turnover intention matter? Evaluating the usefulness of turnover intention rate as a predictor of actual turnover rate. Review of Public Personnel Administration, 36, 240-263.

17.

Cohen

West

S. G.

Aiken

L. S.

(2013). Applied multiple regression/correlation analysis for the behavioral sciences. New York, NY: Routledge.

18.

Cole

M. S.

Bruch

(2006). Organizational identity strength, identification, and commitment and their relationships to turnover intention: Does organizational hierarchy matter? Journal of Organizational Behavior, 27, 585-605.

19.

Coleman

J. S.

(1988). Social capital in the creation of human capital. American Journal of Sociology, 94, S95-S120.

20.

Condrey

S. E.

(2005). Handbook of human resource management in government (2nd ed.). San Francisco, CA: Jossey-Bass.

21.

Cresswell

J. W.

Clark

V. L. P.

(2007). Mixed methods research. Thousand Oaks, CA: SAGE.

22.

Cross

Nohria

Parker

(2002). Six myths about informal networks and how to overcome them. MIT Sloan Management Review, 43, 67-75.

23.

Dalton

D. R.

Johnson

J. L.

Daily

C. M.

(1999). On the use of “intent to . . . ” variables in organizational research: An empirical and cautionary assessment. Human Relations, 52, 1337-1350.

24.

Demerouti

Bakker

A. B.

Nachreiner

Schaufeli

W. B.

(2001). The job demands-resources model of burnout. Journal of Applied Psychology, 86, 499-512.

25.

Dixon

J. C.

(2006). The ties that bind and those that don’t: Toward reconciling group threat and contact theories of prejudice. Social Forces, 84, 2179-2204.

26.

Eberly

M. B.

Holley

E. C.

Johnson

M. D.

Mitchell

T. R.

(2011). Beyond internal and external: A dyadic theory of relational attributions. Academy of Management Review, 36, 731-753.

27.

Ertas

(2015). Turnover intentions and work motivations of millennial employees in federal service. Public Personnel Management, 44, 401-423.

28.

Feeley

T. H.

(2000). Testing a communication network model of employee turnover based on centrality. Journal of Applied Communication Research, 28, 262-277.

29.

Gibbons

D. E.

(2004). Friendship and advice networks in the context of changing professional values. Administrative Science Quarterly, 49, 238-262.

30.

Granovetter

M. S.

(1985). Economic action and social structure: The problem of embeddedness. American Journal of Sociology, 91, 481-510.

31.

Guthrie

(1998). The declining significance of guanxi in China’s economic transition. The China Quarterly, 154, 254-282.

32.

Hitt

Lee

H. U.

Yucel

(2002). The importance of social capital to the management of multinational enterprises: Relational networks among Asian and Western firms. Asia Pacific Journal of Management, 19, 353-372.

33.

Hofstede

(1991). Cultures and organizations: Software of mind. London, England: McGraw-Hill.

34.

Hofstede

(2001). Culture’s consequences: Comparing values, behaviors, institutions, and organizations across nations (2nd ed.). Thousand Oaks, CA: SAGE.

35.

Hogg

M. A.

Terry

D. J.

(2000). Social identity and self-categorization processes in organizational contexts. Academy of Management Review, 25, 121-140.

36.

Huang

T.-C.

Lawler

Lei

C.-Y.

(2007). The effects of quality of work life on commitment and turnover intention. Social Behavior and Personality: An International Journal, 35, 735-750.

37.

Ibarra

(1993). Personal networks of women and minorities in management: A conceptual framework. Academy of Management Review, 18, 56-87.

38.

Campbell

J. W.

Cha

(2013). Revisiting Confucian bureaucracy: Roots of the Korean government’s culture and competitiveness. Public Administration and Development, 33, 286-296.

39.

Kim

(2005). Factors affecting state government information technology employee turnover intentions. The American Review of Public Administration, 35, 137-156.

40.

Krackhardt

Hanson

(1993). Informal networks: The company behind the charts. Harvard Business Review, 71, 104-111.

41.

Krackhardt

Porter

L. W.

(1986). The snowball effect: Turnover embedded in communication networks. Journal of Applied Psychology, 71, 50-55.

42.

Lee

S. Y.

Hong

J. H.

(2011). Does family-friendly policy matter? Testing its impact on turnover and performance. Public Administration Review, 71, 870-879.

43.

Lewis

G. B.

Cho

Y. J.

(2011). The aging of the state government workforce: Trends and implications. The American Review of Public Administration, 41, 48-60.

44.

Long

J. S.

Freese

(2001). Regression model for categorical dependent variables using STATA. College Station, TX: STATA Press.

45.

Marques

J. M.

Yzerbyt

V. Y.

Leyens

J. P.

(1988). The “black sheep effect”: Extremity of judgments towards ingroup members as a function of group identification. European Journal of Social Psychology, 18, 1-16.

46.

McPherson

J. M.

Smith-Lovin

Cook

J. M.

(2001). Birds of a feather: Homophily in social networks. Annual Review of Sociology, 27, 415-444.

47.

Mehra

Kilduff

Brass

D. J.

(1998). At the margins: A distinctiveness approach to the social identity and social networks of underrepresented groups. Academy of Management Journal, 41, 441-452.

48.

Meier

K. J.

Hicklin

(2008). Employee turnover and organizational performance: Testing a hypothesis from classical public administration. Journal of Public Administration Research and Theory, 18, 573-590.

49.

Meyer

J. P.

Allen

N. J.

(1991). A three-component conceptualization of organizational commitment. Human Resource Management Review, 1, 61-89.

50.

Meyer

(2006). Asian management research needs more self-confidence. Asia Pacific Journal of Management, 23, 119-137.

51.

Mitchell

T. R.

Holtom

B. C.

Lee

T. W.

Sablynski

C. J.

Erez

(2001). Why people stay: Using job embeddedness to predict voluntary turnover. Academy of Management Journal, 44, 1102-1121.

52.

Monge

P. R.

Contractor

N. S.

(2003). Theories of communication networks. New York, NY: Oxford University Press.

53.

Mossholder

K. W.

Settoon

R. P.

Henagan

S. C.

(2005). A relational perspective on turnover: Examining structural, attitudinal, and behavioral predictors. Academy of Management Journal, 48, 607-618.

54.

Moynihan

D. P.

Landuyt

(2008). Explaining turnover intention in state government: Examining the roles of gender, life cycle, and loyalty. Review of Public Personnel Administration, 28, 120-143.

55.

Moynihan

D. P.

Pandey

S. K.

(2008). The ties that bind: Social networks, person-organization value fit, and turnover intention. Journal of Public Administration Research and Theory, 18, 205-227.

56.

Nisbett

(2010). The geography of thought: How Asians and Westerners think differently . . . and why. New York, NY: Simon & Schuster.

57.

Chung

M. H.

Labianca

(2004). Group social capital and group effectiveness: The role of informal socializing ties. Academy of Management Journal, 47, 860-875.

58.

O’Reilly

C. A.

III Caldwell

D. F.

Barnett

W. P.

(1989). Work group demography, social integration, and turnover. Administrative Science Quarterly, 34, 21-37.

59.

Polanyi

(1957). The great transformation. Boston, MA: Beacon Press.

60.

Rank

O. N.

(2008). Formal structures and informal networks: Structural analysis in organizations. Scandinavian Journal of Management, 24, 145-161.

61.

Reid

G. B.

Nygren

T. E.

(1988). The subjective workload assessment technique: A scaling procedure for measuring mental workload. In Peter

A. H.

Najmedin

(Eds.), Advances in psychology (Vol. 52, pp. 185-218). North-Holland, The Netherlands: North-Holland.

62.

Robert

Hockey

(1997). Compensatory control in the regulation of human performance under stress and high workload: A cognitive-energetical framework. Biological Psychology, 45(1-3), 73-93.

63.

Saldaña

(2012). The coding manual for qualitative researchers (2nd ed.). Los Angeles, CA: SAGE Publications.

64.

Schalk

Torenvlied

Allen

(2011). Network embeddedness and public agency performance: The strength of strong ties in Dutch higher education. Journal of Public Administration Research and Theory, 20, 629-653.

65.

Selden

S. C.

Moynihan

D. P.

(2000). A model of voluntary turnover in state government. Review of Public Personnel Administration, 20, 63-74.

66.

Siegel

(2007). Contingent political capital and international alliances: Evidence from South Korea. Administrative Science Quarterly, 52, 621-666.

67.

Soltis

S. M.

Agneessens

Sasovova

Labianca

G. J.

(2013). A social network perspective on turnover intentions: The role of distributive justice and social support. Human Resource Management, 52, 561-584.

68.

Sparrowe

R. T.

Liden

R. C.

Wayne

S. J.

Kraimer

M. L.

(2001). Social networks and the performance of individuals and groups. Academy of Management Journal, 44, 316-325.

69.

Steel

R. P.

Lounsbury

J. W.

(2009). Turnover process models: Review and synthesis of a conceptual literature. Human Resource Management Review, 19, 271-282.

70.

Tanova

Holtom

B. C.

(2008). Using job embeddedness factors to explain voluntary turnover in four European countries. The International Journal of Human Resource Management, 19, 1553-1568.

71.

Tichy

N. M.

Tushman

M. L.

Fombrun

(1979). Social network analysis for organizations. Academy of Management Review, 4, 507-519.

72.

Triandis

H. C.

(1995). Individualism and collectivism. Boulder, CO: Westview.

73.

Tsui

A. S.

Egan

T. D.

O’Reilly

C. A.

III . (1992). Being different: Relational demography and organizational attachment. Administrative Science Quarterly, 37, 549-579.

74.

Turner

J. C.

Hogg

M. A.

Oakes

P. J.

Reicher

S. D.

Wetherell

M. S.

(1987). Rediscovering the social group: A self-categorization theory. Oxford, UK: Basil Blackwell.

75.

Uzzi

(1997). Social structure and competition in interfirm networks: The paradox of embeddedness. Administrative Science Quarterly, 42, 35-67.

76.

Van den Bulte

Moenaert

R. K

. (1998). The effects of R&D team co-location on communication patterns among R&D marketing and manufacturing. Management Science, 44, S1-S18.

77.

Westphal

J. D.

Milton

L. P.

(2000). How experience and network ties affect the influence of demographic minorities on corporate boards. Administrative Science Quarterly, 45, 366-398.

78.

Westphal

J. D.

Stern

(2007). Flattery will get you everywhere (Especially if you are a male Caucasian): How ingratiation, boardroom behavior, and demographic minority status affect additional board appointments at U.S. companies. The Academy of Management Journal, 50, 267-288.