Content uploaded by Sang Eun Woo
Author content
All content in this area was uploaded by Sang Eun Woo on Apr 08, 2016
Content may be subject to copyright.
PERSONNEL PSYCHOLOGY
2015, 00, 1–38
INTERNAL AND EXTERNAL NETWORKING
DIFFERENTIALLY PREDICT TURNOVER THROUGH
JOB EMBEDDEDNESS AND JOB OFFERS
CAITLIN M. PORTER
Purdue University
SANG EUN WOO
Purdue University
MICHAEL A. CAMPION
Purdue University
Although the career benefits associated with professional networking
are relatively well established, the repercussions of this highly regarded
career management activity for voluntary turnover have rarely been ex-
amined. Given the potential costs associated with voluntary turnover,
this study sought to clarify the roles of networking behaviors in relation
to voluntary turnover by focusing on the distinction between internal and
external networking. Based on survey responses of industrial and orga-
nizational psychology professionals, we found that internal and external
networking behaviors differentially predicted decisions to voluntarily
leave employers 2 years later: The likelihood of voluntary turnover was
negatively predicted by internal networking and positively predicted by
external networking. Furthermore, to shed light on the reasons why em-
ployee networking behaviors differentially predicted turnover decisions,
this study also examined 4 turnover antecedents (job satisfaction, job
embeddedness, perceived employment opportunities, and job offers) as
potential mediating mechanisms. Results revealed that the relationships
of internal and external networking with voluntary turnover were medi-
ated by job embeddedness and job offers, respectively. We discuss the
implications of these findings for understanding and managing employee
networking and retention.
Although organizational research has recognized the benefits of net-
working for individual career success (Forret & Dougherty, 2004; Wolff
& Moser, 2009), it has rarely considered the implications of this prac-
tice for increasing employees’ chances of leaving their employers (Wolff
& Moser, 2010). Voluntary turnover has the potential to incur substan-
tial costs to organizations both indirectly (e.g., loss of human capital,
Correspondence and requests for reprints should be addressed to Caitlin M. Porter,
Department of Psychological Sciences, Purdue University, 703 Third Street, West Lafayette,
IN 47907; caitlinmporter@gmail.com.
C
2015 Wiley Periodicals, Inc. doi: 10.1111/peps.12121
1
2 PERSONNEL PSYCHOLOGY
organizational knowledge) and directly (e.g., recruitment, selection, train-
ing), which can range from 90% to over 200% of a leaver’s annual salary
(Allen, Bryant, & Vardaman, 2010; Cascio, 2006; Woo & Maertz, 2012).
Yet, few attempts have been made to delineate the conditions under which
employee networking increases the risk of turnover (i.e., when) nor is
there evidence of the psychological process through which networking
affects employees’ decisions to leave (i.e., why). In light of this, this study
answers the questions of when and why employees’ networking behaviors
relate to voluntary turnover.
To address the “when” question, we distinguish between networking
with coworkers within one’s organization (i.e., internal networking)
and networking with colleagues across different organizations (i.e.,
external networking) as two conditions under which networking may
differentially affect the likelihood of voluntary turnover. Prior research
by Wolff and Moser (2010) has laid an important foundation for this
investigation. Based on a sample of working adults in Germany, they
found some support for the suggestion that external (as opposed to
internal) networking positively predicts change of employer 2 years later.
However, only one type of external networking behavior (i.e., maintaining
existing contacts) was positively associated with change of employer,
whereas other types of external networking were either unrelated or
negatively related to change of employer, necessitating further research to
clarify these relationships. At the very least, as suggested by the authors
themselves (Wolff & Moser, 2009, 2010), their findings regarding the
relationships of networking with job mobility within the German context
should be replicated in other national and cultural contexts, considering
that differences in labor markets and economic conditions influence job
mobility (e.g., higher job security and stability in Germany compared
with the United States). Furthermore, our study extends Wolff and
Moser’s theoretical arguments by also considering the potential influence
of internal networking behavior in deterring voluntary turnover.
To address the “why” question, we propose and test a set of psy-
chological mechanisms through which internal and external networking
behaviors may differentially relate to turnover decisions. To do so, we
integrate the emergent body of research on networking (e.g., Forret &
Dougherty, 2001; Wolff & Moser, 2010) with the deep-rooted tradition
of turnover research, which has identified key turnover antecedents (e.g.,
Hom, Caranikas-Walker, Prussia, & Griffeth, 1992; Hom, Mitchell, Lee,
& Griffeth, 2012; Mobley, Griffeth, Hand, & Meglino, 1979; Price &
Mueller, 1981). Specifically, we argue that internal and external net-
working behaviors differentially influence voluntary turnover through
employees’ desires to remain with their employers, which are
closely associated with job satisfaction and job embeddedness, and
PORTER ET AL. 3
opportunities to exit, represented by perceived alternatives and job of-
fers. These turnover antecedents are included in both traditional models
of voluntary turnover (e.g., Mobley et al., 1979; Price & Mueller, 1981) as
well as more contemporary theories of turnover and withdrawal (e.g., Hom
et al., 2012; Mitchell & Lee, 2001), and their predictive validities have
received extensive empirical support (Griffeth, Hom, & Gaertner, 2000;
Jiang, Liu, McKay, Lee, & Mitchell, 2012). Thus, taking into account
networking behaviors with these turnover antecedents affords a more in-
tegrative and informative investigation of the unique roles of networking
behaviors in explaining employee turnover.
Overall, this study makes three contributions to the networking and
turnover literatures. First, it highlights the topic of voluntary turnover as
an important (yet understudied) consequence of employees’ networking
behaviors by explicating when and why employee networking may in-
crease the risk of turnover for organizations. As employees take greater
responsibility for their career management (Arthur & Rousseau, 1996),
knowledge of how proactive career-oriented behaviors (e.g., networking)
influence employees’ work attitudes and behaviors (e.g., job satisfaction,
turnover) is necessary for organizations to develop strategies for effec-
tively managing and retaining their employees. Second, this study clarifies
the respective roles of internal and external networking behaviors in pre-
dicting employees’ decisions to stay or leave their employers, which offers
practical insights into how organizations may manage employees’ internal
or external networking behaviors. Third, our research represents one of
the very first attempts to integrate networking and turnover literatures to
propose and test a theory-driven explanation for why internal and exter-
nal networking behaviors may deter or promote voluntary turnover. By
simultaneously testing multiple, theoretically plausible mediating mecha-
nisms, this study sheds light on which commonly acknowledged turnover
antecedents actually account for the predictive effects of networking on
voluntary turnover. Before introducing our theory and hypotheses, we
first clarify what we mean by “networking behaviors” and discuss how
networking is distinct from other relational constructs.
What Are Networking Behaviors?
In this article, we define networking as “behaviors aimed at building,
maintaining, and using informal relationships that possess the (potential)
benefit of facilitating work-related activities of individuals by voluntar-
ily granting access to resources and maximizing common advantages”
(Wolff & Moser, 2009, pp. 196–197). Building on this definition, we
take a psychological perspective on networking in which we consider
networking behaviors to consist of interpersonal interactions that involve
the exchange of resources between networking partners. Resources in this
4 PERSONNEL PSYCHOLOGY
context are best construed as information, socioemotional support (e.g.,
friendship), services, or influence that networking partners may provide
to and receive from one another (Foa, 1971). Drawing from the logic of
social exchange theory (Blau, 1964; Gouldner, 1960), we suggest that the
continual exchange of resources (via networking behaviors) between two
or more networking partners has the potential to advance each partner’s
work-related goals and to develop a professional network relationship. We
also note that all networking behaviors (or networkers) are not equally ef-
fective: Networking behaviors may be performed differently or exchange
different resources depending on the characteristics of the networking
partners and the stage of their network relationship (Porter & Woo, 2015).
More importantly, network contacts offer access to different resources
depending on whether they are internal or external to one’s employing
organization (e.g., internal contacts more often offer information about
organizational culture whereas external contacts more often offer infor-
mation about best practices in the field). Therefore, the distinction between
internal and external networking is important to consider, as it is the dif-
ferent resources accessible through networking with each set of network
contacts that drives the differential relationships of internal and external
networking with employee outcomes.
To further clarify the networking behaviors construct, we discuss how
networking behaviors differ from other conceptually related relational
constructs—namely, social networks, social capital, coworker support,
and other social effectiveness constructs. We also present common def-
initions and measurement approaches of these constructs in Table 1 to
illustrate how networking behaviors are distinct. First, we distinguish net-
working behaviors from social networks. Social networks refer to the
actual dyadic relationships between an individual and each person with
whom he or she interacts or the structure of relationships among a group of
actors (Tichy, Tushman, & Fombrun, 1979; also see Wasserman & Faust,
1994). Networking behaviors, on the other hand, refer to the activities one
performs to build, maintain, and use such professional relationships. Sec-
ond, we also distinguish networking behaviors from social capital, which
refers to the “goodwill that is engendered by the fabric of social relations
and that can be mobilized to facilitate action” (Adler & Kwon, 2002, p.
17; cf. Coleman, 1988; Lin, 2002; Nahapiet & Ghoshal, 1998). In other
words, social capital refers to the intangible, positive feelings personal
contacts have for one another that facilitate the exchange of interpersonal
resources such as information, influence, and social support (Lin, 2002). In
contrast, networking behaviors are activities directed toward developing
and sustaining multiple professional relationships (or the “social fabric”)
that serve as access points for one to obtain such resources. Thus, net-
working behaviors have the potential to contribute to the development of
PORTER ET AL. 5
TABL E 1
Comparison of Definitions and Measurement Approaches of Networking Behaviors and Related Constructs
Construct Definition Measurement approach
Networking
behaviors
“Behaviors aimed at building,
maintaining, and using informal
relationships that possess the (potential)
benefit of facilitating work-related
activities of individuals by voluntarily
granting access to resources and
maximizing common advantages.”
(Wolff & Moser, 2009; pp. 196–197)
Behavioral Rating Scale (example items):
I use company events to make new contacts. (internal building)
I catch up with colleagues from other departments about what they are
working on. (internal maintaining)
I receive confidential advice in business matters from my contacts in other
departments. (internal using)
I develop informal contacts with professionals outside the organization in
order to have personal links beyond the company. (external building)
For business purposes, I keep in contact with former colleagues. (external
maintaining)
I exchange professional tips and hints with colleagues from other
organizations. (external using)
Social networks The pattern or structure of associations
amongst a set of actors in a network.
Sociometric approach:
Identify existing network contacts and report on relationship and/or contact
characteristics; resulting information is used to create variables that
represent network or relational characteristics (e.g., network size, degree
centrality)
(see Burt, 1992; Scott, 2000; Wasserman & Faust, 1994)
Social capital “Goodwill that is engendered by the fabric
of social relations and that can be
mobilized to facilitate action.” (Adler
& Kwon, 2002, p. 17)
Various approaches including social network measures and subjective
evaluations of social networks and/or resources available
(Continued)
6 PERSONNEL PSYCHOLOGY
TABLE 1 (continued)
Construct Definition Measurement approach
Coworker support Social support received from
organizational members.
Rating Scale (example items):
My coworkers really care about my well-being.
My coworkers care about my general satisfaction at work.
My coworkers show very little concern for me (reverse-scored).
(Mossholder, Settoon, & Henagan, 2005)
Social effectiveness Rating Scale (example items):
Social skill “Interpersonal perceptiveness and the
capacity to adjust one’s behavior to
different situational demands and to
effectively influence and control the
responses of others.” (Ferris, Witt, &
Hochwarter, 2001, p. 1076)
I find it easy to put myself in the position of others.
I am keenly aware of how I am perceived by others.
In social situations, it is always clear to me exactly what to say and do.
I am particularly good at sensing the motivations and hidden agendas of
others.
(Ferris et al., 2001)
Political skill “The ability to effectively understand
others at work, and to use such
knowledge to influence others to act in
ways that enhance one’s personal
and/or organizational objectives.”
(Ahearn, Ferris, Hochwarter, Douglas,
& Ammeter, 2004, p. 311)
I spend a lot of time and effort at work networking with others. (networking
ability)
It is important that people believe I am sincere in what I say and do. (apparent
sincerity)
I always seem to instinctively know the right thing to say or do to influence
others. (social astuteness)
It is easy for me to develop good rapport with most people. (interpersonal
influence)
(Ferris et al., 2005)
PORTER ET AL. 7
social capital (by generating positive feelings between networking part-
ners), but networking behaviors alone are not synonymous with social
capital (Thompson, 2005).
Third, networking behaviors are also distinct from coworker support,
or the social support one receives from other organizational members.
Coworker support may be considered a resource that flows from net-
working interactions with coworkers, as networking behaviors may also
be used to exchange various interpersonal resources (e.g., information).
Thus, coworkers may use networking behaviors to give and receive sup-
port to one another, which may have the result of promoting perceptions
of coworker support (among other perceptions; Kirmeyer & Lin, 1987).
Networking behaviors are further distinguished from coworker support in
that the former can span across internal and external contacts, whereas
coworker support is inherently applicable to relationships within orga-
nizations. Finally, networking behaviors are also distinct from social ef-
fectiveness constructs, such as social and political skill (e.g., Treadway,
Breland, Adams, Duke, & Williams, 2010). These social effectiveness
constructs refer to the talent with which one conducts interpersonal inter-
actions, whereas networking behaviors simply refer to interpersonal inter-
actions with professional colleagues, whether or not they are performed
competently.
Theory and Hypotheses
As a first step toward clarifying “when” networking behaviors pre-
dict voluntary turnover, we integrate our psychological perspective on
networking with contemporary and traditional turnover theories to posit
that internal networking behaviors primarily increase employees’ desires
to remain with their employers, whereas external networking behaviors
primarily increase employees’ opportunities to exit (Hom et al., 2012;
March & Simon, 1958; Mitchell, Holtom, Lee, Sablynski, & Erez, 2001).
Employees’ internal networking behaviors involve positive, mutually ben-
eficial exchanges of interpersonal resources (Wolff & Moser, 2009). If
employees value these interactions or the resources they provide, in-
traorganizational networking behaviors generate constituent forces (i.e.,
attachments to coworkers or organizational groups; Maertz & Griffeth,
2004) that increase their desires to stay at their organizations (unless em-
ployees’ internal contacts prefer to leave; e.g., Felps et al., 2009). On
the other hand, employees who perform external networking behaviors
tend to have more opportunities to exit their organizations as they build
and maintain a professional network of colleagues who are willing or
capable of offering interpersonal resources that are useful for retaining
or acquiring alternative employment (i.e., increasing alternative forces;
8 PERSONNEL PSYCHOLOGY
Griffeth, Steel, Allen, & Bryan, 2005; Hom et al., 2012; Maertz &
Campion, 2004). If employees have more opportunities to leave, they
are likely to exit compared with those who have fewer opportunities for
alternative employment, even if they do not have a strong preference for
leaving (Lee & Mitchell, 1994; Lee, Mitchell, Holtom, McDaniel, & Hill,
1999). Therefore, we propose the following hypotheses:
Hypothesis 1a: Internal networking behaviors will reduce the likeli-
hood of voluntary turnover.
Hypothesis 1b: External networking behaviors will increase the like-
lihood of voluntary turnover.
To answer the question of “why” networking relates to turnover, we
identified four potential mediating variables that are closely associated
with desires to remain (i.e., job satisfaction and job embeddedness) and
opportunities to exit (i.e., perceived employment opportunities and job
offers) based upon prevailing theories of turnover (Hom et al., 2012; Lee
& Mitchell, 1994; Maertz & Campion, 2004; Mitchell & Lee, 2001).
We propose that the negative influence of internal networking behaviors
on voluntary turnover is carried by two complementary mechanisms, job
satisfaction and job embeddedness, which reflect employees’ desires to
remain with their employers (Hom et al., 2012). Specifically, we assert
that employees’ internal networking behaviors increase desires to remain
with their employers by increasing their job satisfaction and embedding
them in their employing organizations. In addition, we argue that the
positive association between external networking and voluntary turnover
is largely explained by its relations with perceived and actual employment
opportunities, which prompt employees’ decisions to leave. We present
an overarching conceptual model in Figure 1, the rationale for which we
discuss in greater detail in the following paragraphs.
To begin, we argue that internal networking behaviors reduce em-
ployees’ likelihood of turnover by increasing their desires to stay, which
is closely linked to job satisfaction (Holtom, Mitchell, Lee, & Eberly,
2008; March & Simon, 1958). Employees’ internal networking behav-
iors increase employees’ perceptions of social support, which, in turn,
increase or sustain their job satisfaction (and reduce their propensity to
exit). More specifically, intraorganizational networking results in the ex-
change of instrumental resources (e.g., advice on how to handle difficult
work tasks) or expressive resources (e.g., emotional support when deal-
ing with a stressful work situation) that increase employees’ perceptions
of coworker support (Kirmeyer & Lin, 1987; Vinokur, Schul, & Caplan,
1987), which improves or maintains their job satisfaction (Chiaburu &
PORTER ET AL. 9
Figure 1: Conceptual Model of Networking Behaviors and Voluntary
Turnover.
Harrison, 2008). When employees are more satisfied with their jobs, they
are less likely to leave for alternative employment (Griffeth et al., 2000).
We also recognize that, just as internal networking may encourage
job satisfaction, job satisfaction may encourage employees to build and
maintain relationships with coworkers to sustain a positive and enjoyable
work environment (Humphrey, Nahrgang, & Morgeson, 2007). Likewise,
job dissatisfaction often encourages employees to withdraw from work
(Hulin, 1991), which may include eschewing networking interactions with
workplace colleagues. Taking these ideas together, internal networking
and job satisfaction may reciprocally influence one another. However,
we maintain that internal networking behaviors are likely to exhibit a
stronger predictive relationship with job satisfaction (as opposed to the
reverse causal ordering) because networking behaviors (in general) are
primarily driven by the desire to acquire valued interpersonal resources
that facilitate work-related activities (Wolff & Moser, 2009). Compared
with this impetus, job satisfaction likely plays a lesser role in encouraging
internal networking behaviors. Thus, we hypothesize the following:
Hypothesis 2: Job satisfaction partially mediates the relationship
between internal networking and voluntary turnover.
10 PERSONNEL PSYCHOLOGY
Another means by which intraorganizational networking behaviors de-
crease the likelihood of voluntary turnover is through fostering employees’
job embeddedness. Defined as “the extent to which people feel attached,
regardless of why they feel that way, how much they like it, or whether
they chose to be so attached” (Crossley, Bennett, Jex, & Burnfield, 2007,
p. 1032), job embeddedness refers to employees’ psychological attach-
ments to their jobs/organizations that arise from three components: links,
fit, and sacrifice. Links refer to the connections to people or groups; fit
refers to the extent to which one’s job fits with other aspects of one’s life;
and sacrifice refers to the relationships and activities that one would lose
by leaving his or her employer (Mitchell et al., 2001).
Drawing from job embeddedness theory, we suggest three ways in
which internal networking protects against turnover by engendering psy-
chological attachments to the job or organization (even in the presence
of job satisfaction). First, networking behaviors lead to the development
of coworker relationships, which represent links within the organization
that produce affective (i.e., positive emotional reactions to the organi-
zation), constituent (i.e., attachments to coworkers), and normative (i.e.,
pressure to stay from friends) motivations to remain with the organization
(Maertz & Campion, 2004; Mitchell et al., 2001). Second, the exchange of
interpersonal resources via networking helps employees develop greater
comfort with work-related tasks and their social environments, which in-
creases perceptions of “fit” with the organization (Cable & Parsons, 2001;
Morrison, 2002). Finally, the links employees develop through internal
networking behaviors represent sacrifices in the form of psychological
costs associated with leaving the organization and thus serve as moti-
vational forces that encourage individuals to stay (Maertz & Griffeth,
2004). Taking these ideas together, it is likely that internal networking
behaviors encourage employees to remain with their employers via job
embeddedness.
At the same time, prior research has shown that job embeddedness
actually predicts a decrease in the rate of activities directed toward build-
ing social capital within the organization, which are akin to networking
behaviors (Ng & Feldman, 2010). This finding illustrates that employee
networking behaviors do not exist in static isolation but are dependent
upon the needs of the employee. Once employees have established “links”
at their employer (i.e., are embedded), they may have less of a need to
continue building relationships, and they may shift their networking ef-
forts toward maintaining and using their established network relation-
ships so as to more easily access valued resources. Based on this ratio-
nale, we argue that internal networking behaviors not only promote job
embeddedness but also sustain embeddedness after employees perceive
themselves to be embedded. Thus, we propose the following hypothesis:
PORTER ET AL. 11
Hypothesis 3: Job embeddedness partially mediates the relationship
between internal networking and voluntary turnover.
Finally, we assert that employees’ external networking behaviors in-
crease their likelihood of voluntary turnover by creating more opportu-
nities to exit their organizations for alternative employment, which we
represent as both perceived and actual (i.e., job offers) employment op-
portunities. We argue that employees’ external networking increases their
perceived employment opportunities by providing access to information
about alternative job opportunities (i.e., informational resources). Indeed,
research has shown that extraorganizational networking often allows the
exchange of job-related information across firms (Van Hoye & Lievens,
2009), which leads to a greater awareness of alternative employment op-
portunities (Griffeth et al., 2005; Steel, 2002). The reverse may also be
true; perceived employment opportunities may encourage individuals to
engage in external networking behaviors in order to maintain alternative
employment options or validate their perceptions of their employment op-
portunities. Despite this possibility, we maintain that external networking
behaviors are an initial determinant of perceived employment opportuni-
ties, as networking is a fundamental career competency that contributes to
people’s perceptions of their abilities to find and retain employment (Eby,
Butts, & Lockwood, 2003). With increased perceptions of alternative em-
ployment opportunities, employees’ propensities to exit their employing
organizations also increase (March & Simon, 1958; Mobley et al., 1979).
We also propose that employees’ external networking behaviors in-
crease the likelihood of voluntary turnover because they contribute to the
attainment of job offers (Granovetter, 1995). In fact, external networking
has been posited as an effective job search strategy (e.g., Wanberg, Kanfer,
& Banas, 2000), and networking behaviors focused on job search activi-
ties have been shown to positively predict the number of subsequent job
offers received (Van Hoye, Hooft, & Lievens, 2009). In addition, just as
those who network with extraorganizational colleagues receive informa-
tion about alternative job opportunities, they may also provide information
about their competencies or desire for movement to their external contacts.
In turn, employees’ external contacts may use their influence with third-
party contacts (i.e., a friend of a friend) to recommend the employee for a
job. Indeed, research has shown that individuals often acquire jobs through
third-party contacts (Bian, 1997) and that individuals are more likely to
be offered a job when a friend refers them (Fernandez & Weinberg, 1997).
Once a job offer is in hand, one is likely to accept it depending upon its
attractiveness and perceived usefulness (Griffeth et al., 2000; Mobley
et al., 1979). Based on these arguments, we propose the following
hypotheses:
12 PERSONNEL PSYCHOLOGY
Hypothesis 4a: Perceived employment opportunities partially medi-
ate the relationship between external networking and
voluntary turnover.
Hypothesis 4b: Job offers partially mediate the relationship between
external networking and voluntary turnover.
Method
Participants and Procedure
To increase the probability that the study sample participated in both
external networking and internal networking behaviors, we contacted
members of a professional association (Society for Industrial and Or-
ganizational Psychology or SIOP) registered in a professional recruitment
database to participate in two waves of data collection via online survey.
The first survey included measures of demographic information, network-
ing behaviors, job satisfaction, job embeddedness, perceived employment
opportunity, job search, and job offers. Of the 2,936 professionals con-
tacted, 540 completed the first survey for a response rate of about 18.4%.
From this initial sample, we selected 371 individuals who had been work-
ing in applied settings for at least 2 years. The remaining respondents
either were not working in applied settings or had left their employing
organization within the past 2 years. This 2-year time frame allowed us to
control for potential socialization effects and changes in work attitudes that
are commonly found in early employment stages, which could potentially
affect turnover decisions in addition to networking efforts (e.g., Boswell,
Boudreau, & Tichy, 2005; Fisher, 1986). Two years later (Time 2), we
requested that these respondents fill out a second online survey, and 266
participants responded (a continued response rate of 72%). The second
survey assessed whether participants had voluntarily left their employers
in the past 2 years.
Both nonresponse bias and attrition bias could potentially influence
the validity of study findings. To investigate nonresponse bias, we com-
pared early survey responders with late survey responders, who may be
considered a proxy for nonrespondents (Rogelberg & Stanton, 2007).
We found no significant differences between early and late responders on
study variables, suggesting that nonresponse bias is not present in our data.
To evaluate the potential for attrition bias, we followed the recommenda-
tions of Goodman and Blum (1996). First, we regressed a dichotomous
indicator of participant attrition at Time 2 (0 =left the study;1=stayed
in the study) on our study variables using logistic regression; we found
that participants were more likely to remain in the study when they were
less embedded in their jobs (B=−.49, p=.07, OR =.61), were more
PORTER ET AL. 13
satisfied with their jobs (B=.44, p=.06, OR =1.55), and did not per-
ceive alternative employment opportunities (B=−.18, p=.08, OR =
.83). These findings indicate that nonrandom sampling is present at Time
2. Second, we examined whether the study variable means of Time 2
respondents differed from nonrespondents. Compared to nonrespondents,
Time 2 respondents were less likely to perform internal networking be-
haviors (t(344) =−2.12, p<.05), less likely to perceive that they had
alternative employment opportunities available to them (t(359) =−2.20,
p<.05), and less likely to search for alternative employment (t(365) =
−2.71, p<.05). Finally, we compared the variances of study variables for
the full sample (i.e., Time 1) and the final sample (i.e., Time 2) to evaluate
whether attrition influenced variances of study variables, and we found
that the variance of job search was restricted at Time 2 (z=−2.46, p<
.01). Taken together, these findings reveal that attrition may bias study
findings; we discuss the implications of attrition bias for this study in the
Discussion section.
Measures
Below, we describe how the criterion variable (i.e., voluntary turnover)
was measured followed by a list of predictors included in the study. All
measures were assessed using a 1 =strongly disagree to 5 =strongly
agree scale unless otherwise noted.
Voluntary turnover. Voluntary turnover was assessed at Time 2 by
asking participants if they had voluntarily left their employer within the
past 2 years. If they answered “Yes,” they were classified as voluntary
leavers (n=25).
Networking behaviors. In order to measure internal and external net-
working behaviors, we used 18 items selected from the 44-item scale
developed by Wolff and colleagues (Wolff & Moser, 2006; Wolff,
Schneider-Rahm, & Forret, 2011). The original 44-item scale is based
upon two structural dimensions of internal and external networking and
three functional dimensions of building,maintaining, and using contacts.
Each item represents a networking behavior that captures one structural
dimension and one functional dimension. Participants rated the frequency
with which they engaged in each behavior on a four-point scale ranging
from 1 =never/very seldom to 4 =very often/always. We used vari-
ous item analysis techniques to identify 18 items that best represented
the constructs of internal and external networking behaviors—nine items
for each construct. Details of the item selection procedure and psycho-
metric properties of the reduced scale are reported in Appendix A. To
account for the underlying theory of the networking behaviors measure,
we adopted a measurement model that contains two factors, internal and
14 PERSONNEL PSYCHOLOGY
external networking, each with three indicators—building,maintaining,
and using scale scores, which showed acceptable fit (χ2(8, N=349) =
36.29, p<.001, Comparative Fit Index [CFI] =.97, Tucker-Lewis Index
[TLI] =.95, Root Mean Square Error of Approximation [RMSEA] =.10,
Standardized Root Mean Square Residual [SRMR] =.05). We provide
a more detailed rationale for the theoretical distinctions between internal
and external networking behaviors in Appendix B as suggested by the
action editor and an anonymous reviewer.
Job satisfaction. Job satisfaction was measured with the eight-item
Abridged Job in General scale (Russell et al., 2004). Participants rated
how “good” or “undesirable” their jobs were.
Job embeddedness. Participants’ job embeddedness was assessed us-
ing Crossley et al.’s (2007) seven-item global job embeddedness scale,
which measures individuals’ attachment to their employers, regardless of
the reasons they are attached. An example item is, “I feel tied to this
organization.”
Perceived employment opportunity. Perceived employment opportunity
was measured using the three-item “Ease of Movement” subscale from
the Employment Opportunity Index which captures perceived availability
of job alternatives on a 1 =strongly disagree to 7 =strongly agree scale
(Griffeth et al., 2005). A sample item is, “Given my qualifications and
experience, getting a new job would not be very hard at all.”
Job offers. Whether or not participants received one or more job offers
from alternative employers was assessed using the following question:
“Within the past 12 months, have you received one or more job offers?”
(1 =yes,0=no). Sixty-six participants (19%) indicated that they had
received at least one job offer.
Control variables. When predicting voluntary turnover, we controlled
for age and organizational tenure because they have been shown to relate
to voluntary turnover (Griffeth et al., 2000), and we controlled for firm
size and education (0 =master’s degree; 1=professional or doctorate
degree) because they were related to networking behaviors in this sample.
We assessed firm size using the following scale: 1 (1–50), 2 (51–100), 3
(101–250), 4 (251–500), 5 (501–1,000), 6 (1,001–3,000), 7 (3,001 or more
employees). Age, organizational tenure, and firm size variables were trans-
formed using either the square root or inverse transformation to ensure that
the variable distributions approximated normality. We also controlled for
job search effort because researchers have suggested that networking may
be used as a job search strategy (e.g., Wanberg et al., 2000), and job search
is a recognized turnover antecedent (Blau, 1993; Lee, Gerhart, Weller, &
Trevor, 2008). Thus, controlling for job search provides more conceptual
clarity in understanding the unique role of networking in predicting study
outcomes. We measured job search effort using Blau’s (1993) four-item
PORTER ET AL. 15
measure in which participants rated items about the time and effort they
devoted to job search during the prior 12 months. A sample item is, “Spent
a lot of time looking for a job alternative.”
Evaluating Common Method Variance and Discriminant Validity
We attempted to mitigate the influence of common method variance
through our study design by assessing the dependent variable at a different
time point. However, as all Time 1 variables were assessed using only
one method (self-report), common method variance (CMV) may pose a
threat to the empirical relationships between networking behaviors and job
satisfaction, perceived employment opportunities, and job embeddedness.
We used the unmeasured latent method factor (ULMF) approach to as-
sessing CMV, which is commonly used in studies with a similar research
design as this study (Podsakoff, MacKenzie, Lee, & Podsakoff, 2003).
We used confirmatory factor analytic techniques to compare a baseline
model containing variables assessed using the self-report method at Time
1 (i.e., external networking, internal networking, job satisfaction, job em-
beddedness, perceived employment opportunities, and job search effort)
with a ULMF model that includes the Time 1 variables as well as an un-
measured latent method factor (that was uncorrelated with the other latent
variables). The ULMF model did not converge because of problems with
identification. As an alternative model, we set all of the factor loadings of
the method factor to equality, which has been adopted by some researchers
as a means for achieving a convergent solution in these circumstances (cf.
Podsakoff et al., 2003). The change in χ2statistic (χ2(2, N=370) =
19.87, p<.001) revealed that the alternative ULMF model (χ2(723, N=
370) =1,539.74, p<.001; CFI =.92; TLI =.91; RMSEA =.06; SRMR
=.06) exhibited better fit compared to the baseline model (χ2(725, N=
370) =1,559.61, p<.001; CFI =.92; TLI =.91; RMSEA =.06; SRMR
=.06), which suggests that CMV is present in our data.
Similar to previous research (e.g., Boswell, Olson-Buchanan, & Har-
ris, 2014), we conducted additional analyses, which suggested that CMV
was unlikely to influence study results. First, we found that the variance
attributable to an ULMF (8%) was below the amount typically found in
similar studies (Carlson & Kacmar, 2000; Williams, Cote, & Buckley,
1989). Second, we found that the differences between the standardized
factor loadings of the baseline and ULMF models were minimal (average
difference was .058; largest difference was .126) and comparable to
previous research (Boswell et al., 2014). Finally, we found that the
RMSEA values of each model had overlapping confidence intervals,
providing evidence that the baseline model is more parsimonious. Taking
these findings together, we concluded that common method variance
posed a limited threat to our study results. Nonetheless, we note that this
16 PERSONNEL PSYCHOLOGY
procedure only accounts for CMV attributable to a common correlation
across all variables within a single method and does not address
CMV caused by other factors, such as respondents’ active information
processing and self-generated validity (Feldman & Lynch, 1988).
In addition, we investigated distinctiveness of the six variables mea-
sured at Time 1 by running a series of confirmatory factor analytic models
(Brown, 2006). The results of these analyses are reported in Table 2.
We found that a six-factor model, where all Time 1 variables were mod-
eled separately, fit better than theoretically plausible alternative models,
which provides some evidence that the Time 1 variables are empirically
distinguishable.
Analytic Strategy
To test Hypotheses 1–4, we used a structural equation modeling (SEM)
approach with a weighted least squares mean and variance adjusted es-
timation procedure in Mplus version 7.1, which converts a dichotomous
dependent variable (i.e., voluntary turnover) to a latent continuous vari-
able while simultaneously estimating the relationships among a set of
variables of interest (e.g., multiple independent and mediating variables).
In our specific case, this approach enabled us to evaluate the respec-
tive roles of internal and external networking behaviors in the turnover
process by simultaneously examining associations among networking,
turnover antecedents, and turnover as opposed to examining these links in
artificial isolation. Moreover, this approach is advantageous over logistic
regression (which is often used to analyze dichotomous outcomes) be-
cause it allows for the inclusion of latent internal and external networking
behaviors variables that incorporate building, maintaining, and using net-
working behaviors scale scores as indicators. One potential drawback of
this approach is that it does not yield odds ratios, which are a useful metric
for evaluating the likelihood of a dichotomous event (e.g., turnover) when
a predictor is of a certain value. Thus, when investigating Hypothesis 1,
we calculated the odds ratio for each predictor by dividing the odds of
turnover when there is a one standard deviation increase in a study vari-
able with all other variables held constant at their means by the odds of
turnover when all variables are held constant at their means.1
1We calculated the odds for each variable using the following equation: P=1 – proba-
bility ([threshold – z]/ϴ)), where threshold =the threshold of the dichotomous event, ϴ
=the residual variance of the dichotomous event, and z=a*η1+b*η2 . . . n*x, where a,
b,...n=the unstandardized regression coefficient associated with a particular variable;
η1, η2...ηN=the level of the variable (in this case, we chose the mean for all variables);
x=the level of the variable being evaluated (in our case, one standard deviation above the
mean; Muth´
en, 2005).
PORTER ET AL. 17
TABL E 2
Comparison of Measurement Models for Study Variables
Model Model description χ2df χ2df CFI TLI RMSEA SRMR
Baseline
six-factor
model
Internal networking, external networking, job
satisfaction, job embeddedness, perceived
employment opportunity, and job search effort
as separate factors
1559.61 725 .92 .91 .06 .06
Model 1 Five factors: internal networking and external
networking combined into one factor
2158.28 730 598.67∗∗ 5 .86 .85 .07 .08
Model 2 Five factors: job satisfaction and job
embeddedness combined into one factor
2132.19 730 572.58∗∗ 5 .86 .85 .07 .06
Model 3 Five factors: employment opportunity and job
search effort combined into one factor
2031.34 730 471.73∗∗ 5 .87 .86 .07 .08
Model 4 Four factors: internal networking, job
satisfaction, and job embeddedness combined
into one factor
3189.10 734 1629.49∗∗ 9 .76 .74 .10 .12
Model 5 Four factors: external networking, perceived
employment opportunity, and job search effort
combined into one factor
4256.02 734 2696.41∗∗ 9 .65 .63 .11 .14
Model 6 Two factors: internal networking, job
satisfaction, and job embeddedness combined
into one factor and external networking,
perceived employment opportunity, and job
search effort combined into one factor
5858.66 739 4299.05∗∗ 14 .49 .46 .14 .17
Model 7 All items combined into one factor 6844.77 740 5285.16∗∗ 15 .39 .36 .15 .17
Note. CFI =Comparative Fit Index; TLI =Tucker-Lewis Index; RMSEA =Root Mean Square Error of Approximation; SRMR =Standardized Root
Mean Square Residual.
∗∗ p<.01.
18 PERSONNEL PSYCHOLOGY
We tested the hypothesized model with job search effort included as a
statistical control variable; this model fit the data well (χ2(43, N=365)
=53.71, p=.13; CFI =.98; TLI =.98; RMSEA =.03). To account
for the measurement error in our study variables, we used factor scores
based on regression to approximate the underlying factors of the turnover
antecedents (i.e., job satisfaction, job embeddedness, perceived employ-
ment opportunities, and job search effort). The factor scores represent
the shared variance between the underlying factor and the standardized,
observed items weighted by regression coefficients; they are standardized
with a mean of zero and a standard deviation of one (DiStefano, Zhu,
&M
ˆ
ındrilˇ
a, 2009; Grice, 2001). When specifying the structural model,
we allowed internal networking to correlate with perceived employment
opportunities to account for the associations between internal networking
and perceived marketability, which is closely associated with perceived
employment opportunities (Eby et al., 2003). In addition, in accordance
with turnover theories (e.g., Hom et al., 1992; Mobley et al., 1979), we al-
lowed job satisfaction to correlate with job embeddedness, and perceived
employment opportunities to correlate with job offers. When we included
demographic variables (i.e., age, education, tenure, and firm size) in the
SEM model as statistical controls, they were not associated with voluntary
turnover at statistically significant levels, and they did not alter the sta-
tistical significance of the substantive results. Therefore, we retained and
interpreted a structural model without demographic variables included.
For Hypotheses 2–4, we incorporated the bootstrapping procedure
with 5,000 samples and 90% bias-corrected confidence intervals (CI) to
test the indirect effects of networking behaviors on voluntary turnover
(MacKinnon, Lockwood, & Williams, 2004). The bootstrapping proce-
dure resamples the data with replacement to generate a distribution that is
used to estimate model parameters. This technique is well suited for test-
ing mediation because it has higher power to detect effects (the parameter
estimates are derived from a distribution based on 5,000 samples), and
it is less sensitive to nonnormal distributions (MacKinnon, Lockwood,
Hoffman, West, & Sheets, 2002; Preacher & Hayes, 2008).
Results
The means, standard deviations, and intercorrelations of study vari-
ables are shown in Table 3. To examine the relationships between internal
and external networking behaviors and voluntary turnover (Hypotheses 1a
and 1b), we tested a path model (with a measurement component included
for both internal and external networking behaviors) where internal and
external networking behaviors were directly related to voluntary turnover;
path estimates and odds ratios of study variables are shown in Table 4.
PORTER ET AL. 19
TABL E 3
Means, Standard Deviations, Intercorrelations of Study Variables
Vari a bl e Mean SD 1234567891011
Time 1
1. Age (sqrt) 6.37 .76 –
2. Education .67 .47 .37∗∗ –
3. Tenure (sqrt) 2.37 1.11 .43∗∗ .16∗∗ –
4. Firm size (inv) .46 .36 .01 −.05 .04 –
5. Job search efforta.02 1.00 −.07 −.12∗−.26∗∗ −.07
6. Internal networkingb– .52 .10 .07 .09 .03 −.01 –
7. External networkingb–.55.15
∗∗ .22∗∗ −.03 .04 −.04 .52∗∗ –
8. Job satisfactiona−.10 .98 .16∗∗ .14∗∗ .07 .08 −.37∗∗ .30∗∗ .17∗∗ –
9. Job embeddednessa−.14 .92 .14∗∗ .14∗∗ .31∗∗ .16∗∗ −.43∗∗ .32∗∗ .16∗∗ .69∗∗ –
10. Perceived employment opportunitya−.01 .92 −.01 .08 .07 −.06 −.21∗∗ .22∗∗ .27∗∗ .24∗∗ .19∗∗ –
11. Job offer .18 .38 .03 .08 −.12∗.02 .17∗∗ .19∗.30∗∗ −.06 −.12∗.21∗∗ –
Time 2
12. Voluntary turnover .09 .29 −.15∗−.10 −.15∗−.07 .29∗∗ −.16 .09 −.19∗∗ −.30∗∗ −.04 .24∗∗
Note. aIndicates a standardized factor score. bIndicates a latent variable. Sqrt and Inv indicate that the variables were transformed using the square root
or the inverse transformation, respectively.
∗p<.05, ∗∗ p<.01.
20 PERSONNEL PSYCHOLOGY
TABL E 4
Path Estimates of Direct Relationships of Networking Behaviors on Voluntary
Turnover
BSEβOR R2
Control variables .20
Age (sqrt) −.31 .28 −.21 .58
Education −.06 .34 −.03 .94
Tenure (sqrt) −.02 .18 −.01 .96
Firm size (inv) −.30 .36 −.10 .79
Job search efforta.41∗∗ .12 .32 2.40
Networking behaviors .09
Internal −.78∗∗ .30 −.36 .38
External .64∗∗ .24 .32 2.14
Tot al R2.29
Note. Turnover coded as 1, other responses coded as 0. aIndicates a standardized factor
score. Sqrt and Inv indicate that the variables were transformed using the square root or the
inverse transformation, respectively.
∗∗p<.01. ∗p<.05.
Hypothesis 1a proposed that internal networking behaviors would reduce
the likelihood of voluntary turnover. This hypothesis was supported:
Internal networking behaviors were associated with a reduction in
the likelihood of voluntary turnover. To evaluate the extent to which
internal networking behaviors contribute to the prediction of turnover,
we calculated the odds of turnover for a one standard deviation decrease
in internal networking behaviors when the remaining predictors are held
constant at their respective means. With all else held equal, reducing the
frequency of internal networking behaviors by one standard deviation
was associated with a 2.40 times increase in the likelihood of voluntary
leaving (OR =2.40). In other words, reducing internal networking by
one standard deviation was associated with a 140% increase in the
likelihood of turnover, ceteris paribus. Hypothesis 1b proposed that
external networking would increase the likelihood of voluntary turnover.
This hypothesis was also supported: External networking behaviors
were associated with a rise in the likelihood of voluntary turnover. More
specifically, holding the other predictors constant at their means, a one
standard deviation increase in the frequency of external networking
behaviors was associated with a 2.14 times increase in the likelihood of
voluntary turnover or a 114% increase in the odds of turnover.
Hypotheses 2 and 3 proposed that the negative association between
internal networking behaviors and voluntary turnover would be partially
mediated by job satisfaction and job embeddedness, respectively. Al-
though we did not find support for Hypothesis 2 (B=.05, 90% CI [−.20,
.32]), we did find support for Hypothesis 3; job embeddedness accounted
for the relationship between internal networking behaviors and voluntary
PORTER ET AL. 21
Figure 2: Unstandardized and Standardized Path Estimates for the
Hypothesized Model of Networking and Voluntary Turnover.
Note. The unstandardized path estimates are denoted by the capital B, and the standardized path
estimates are denoted by the beta symbol.
Statistical control: Job search effort.
**p<.01.
turnover, as evidenced by a 90% confidence interval that did not contain
zero (B=−.40, 90% CI [−.74, −.14]). We also note that internal net-
working behaviors were positively associated with both job satisfaction
and job embeddedness, but only job embeddedness exhibited a significant
negative relationship with voluntary turnover (see Figure 2).
Finally, we proposed that the positive association between external
networking behaviors and the likelihood of voluntary turnover would
be partially mediated by perceived employment opportunities (Hypothe-
ses 4a) and job offers (Hypotheses 4b). Although we did not find
support for the mediating role of perceived employment opportunities
(B=−.11, 90% CI [−.38, .05]), we did find that job offers partially me-
diated the relationship between external networking behaviors and volun-
tary turnover (B=.39, 90% CI [.14, .95]). Moreover, consistent with our
theoretical rationale, external networking behaviors were positively asso-
ciated with perceived employment opportunity and job offers. However,
only job offers were positively associated with voluntary turnover.
Supplementary Analyses
The research design of this study leaves open the potential for reverse
causality because the independent variables (i.e., networking behaviors)
22 PERSONNEL PSYCHOLOGY
and mediators (i.e., job satisfaction, job embeddedness, and perceived
employment opportunity) were measured concurrently. Although we ad-
dressed this possibility logically when introducing the rationales for these
hypotheses and empirically by testing for common method variance, we
also examined the potential for reverse causality by evaluating five alter-
native models where job satisfaction and job embeddedness are positioned
as predictors of internal networking and perceived employment opportu-
nities is positioned as a predictor of external networking behaviors both
separately and simultaneously. Both the hypothesized and the alternative
models fit the data similarly, leaving open the possibility of reverse causal-
ity. We discuss the implications of reverse causality for our findings and
offer directions for future research to shed light on the causal ordering of
these variables in the Discussion section.
Discussion
In an effort to shed light on when and why employee networking behav-
iors influence voluntary turnover, this study examined the direct and indi-
rect associations between two types of employee networking behaviors—
internal and external networking—and voluntary turnover. Specifically,
this study clarified how internal and external networking behaviors play
distinct roles in deterring and facilitating voluntary turnover processes by
linking employees’ internal and external networking behaviors to their
turnover decisions 2 years later and testing four turnover antecedents as
potential mediating mechanisms of the networking–voluntary turnover
relationships. Below, we address how this study contributes to the bur-
geoning literature on networking and to the more mature body of research
on voluntary turnover.
To begin, this study accomplishes the goal of clarifying how differ-
ent types of networking behaviors relate to voluntary turnover. Although
previous research has yielded indefinite findings (i.e., Wolff & Moser,
2010), this study reveals that internal networking behaviors are associated
with a reduced likelihood of voluntary turnover, and external network-
ing behaviors are associated with an increased likelihood of voluntary
turnover. Based on these findings, this study indicates that employees’
internal and external networking behaviors have very different organiza-
tional implications. Indeed, employee networking, in general, functions
as a double-edged sword by simultaneously exerting opposing influences
upon one’s desire and ability to leave the organization. On the one hand,
employees’ internal networking behaviors tended to increase their de-
sires to remain, which is closely associated with job satisfaction and job
embeddedness. On the other hand, employees’ external networking be-
haviors tended to increase their opportunities to leave. Together, these
PORTER ET AL. 23
results suggest that whether one is networking with colleagues within or
outside of the organization is a critical distinction because each type of
networking behavior promotes access to different interpersonal resources
and therefore is uniquely associated with voluntary turnover antecedents
and decisions.
Drawing from both traditional and contemporary turnover models,
this study also offers a theoretical explanation and empirical evidence for
why internal and external networking behaviors differentially influence
voluntary turnover decisions. Extending previous research that has inves-
tigated direct relationships between networking and turnover (Wolff &
Moser, 2010), study findings suggest that employees’ internal networking
behaviors reduce their propensity to voluntarily leave by increasing their
desires to remain with their employers (represented by job embedded-
ness), whereas employees’ external networking behaviors increase their
likelihood of turnover by increasing their opportunities to exit (repre-
sented by job offers). By simultaneously investigating multiple mediating
mechanisms, this study helps address the question of why internal and
external networking behaviors relate to voluntary turnover: through in-
creasing desires to remain and opportunities to leave.
Finally, we present empirical evidence that contributes to researchers’
limited knowledge of the antecedents of job embeddedness (cf. Holtom,
Mitchell, & Lee, 2006; Lee, Burch, & Mitchell, 2014). Extant job embed-
dedness research offers few suggestions for what predicts job embedded-
ness (e.g., leader–member exchange; Harris, Wheeler, & Kacmar, 2011).
This study offers some insight on this matter: Our findings suggest that
when employees build, maintain, and use their professional relationships
within their employing organizations, they are likely to feel embedded
in their organizations. Rather than addressing external factors that act
on individuals’ decisions to leave, we identify a proactive behavior (i.e.,
networking) within employees’ control that is associated with whether
or not they prefer to stay. As such, this is one of the few studies that
investigates how employees can take an active role in shaping their reac-
tions toward their work surroundings and, therefore, their propensity to
turnover (Judge, Hulin, & Dalal, 2011).
Practical Implications
The findings of this study also provide insights into whether and
how organizations (and individuals) may promote or manage employee
networking behaviors and voluntary turnover. First, our findings suggest
that employee networking has both positive and negative implications
for organizational functioning: Employees’ external networking may
lead to higher turnover rates, whereas internal networking tends to be
24 PERSONNEL PSYCHOLOGY
more beneficial for the organization (higher job satisfaction and job
embeddedness and lower likelihood of turnover). In light of these results,
organizations may consider the frequency of networking behaviors to be
an important indicator of employees’ inclinations to turnover. External
networking in and of itself does not necessarily indicate that employees
desire or plan to leave, but extreme external networking (i.e., one standard
deviation above average) may indicate an increased propensity to exit.
Likewise, internal networking behaviors at extremely low levels may
indicate employees’ impending job withdrawal. Thus, organizations
may consider identifying, monitoring, and managing employees who
engage in extensive external networking behavior (relative to internal
networking behavior) or limited internal networking behaviors, which
may translate to substantial monetary savings depending on the positions
of the employees (e.g., entry vs. mid-level), employees’ performance
levels, and the size of the organization (Allen et al., 2010; Cascio, 2006).
Moreover, these results suggest that organizations may need to more
seriously consider not only whether employees network, but also with
whom they network. As such, organizations should consider investing
more resources in creating networking opportunities among their own
employees (e.g., online community for the organization, mentoring pro-
grams, company events). As an example, Forret (2013) suggested that
organizations implement cross-functional teams, project groups, and job
rotation to facilitate the development of extensive intraorganizational net-
works. Such internal networking opportunities may increase positive work
experiences, which contribute to positive workplace evaluations (i.e., job
satisfaction) and promote psychological attachments (i.e., job embedded-
ness; Holtom et al., 2006). As such, internal networking may not only
mitigate turnover (by creating reasons to stay) but also improve employ-
ees’ experience of work (e.g., Weiss & Cropanzano, 1996).
On a related note, given that job embeddedness explained why em-
ployee internal networking behaviors reduced the likelihood of voluntary
turnover, organizations may also identify human resource management
practices that increase employees’ inducements to remain with the orga-
nization (e.g., generous benefit packages; Batt & Colvin, 2011). Along
these lines, Holtom et al. (2006) offer other suggestions for increasing job
embeddedness, such as promoting from within (associated with fit), spon-
soring social events (associated with links), and providing on-site services
(associated with sacrifice). Moreover, given that job offers mediated the
relationship between external networking and voluntary turnover, human
resources representatives may consider ways of counteracting employees’
offers for alternative employment by developing a system for posing com-
petitive counteroffers or investing in competitive retirement investments
or securities (e.g., 401K). Finally, our results demonstrate the benefits of
PORTER ET AL. 25
networking for one’s perceived and actual ability to get a (better) job,
which may behoove vocational and career counselors who may consider
recommending networking as an effective practice to increase job search
self-efficacy and effectiveness (Boswell, Zimmerman, & Swider, 2012;
Van Hoye et al., 2009).
Limitations and Future Directions
The current research design has some limitations that should be con-
sidered when interpreting the results of this study. First, we collected data
across two time points in order to reduce problems of common method
variance, but the study is still correlational and the results should be in-
terpreted accordingly. First and foremost, the current research design is
unable to provide conclusive evidence of the causal relationship between
networking and the proposed mediators, as they were measured concur-
rently. Although there is compelling theoretical evidence in support of
the direction of the proposed empirical relationships as discussed earlier
in this article, it is also possible that job satisfaction and job embedded-
ness promote one’s efforts toward internal networking and that perceived
employment opportunities lead to more active engagement in external net-
working behaviors. Therefore, we encourage future research efforts with a
more rigorous, longitudinal design—that is, one that measures networking
behaviors and the proposed mediators both concurrently and repeatedly
(at least three times; Ployhart & Ward, 2011), which would allow the
comparison of the hypothesized causal effects of networking behaviors
on the theorized mediators with the reversed causal effects.
Second, the time-separated design of this study also opens up the pos-
sibility for attrition bias to influence study findings. Follow-up analyses
revealed that participants who remained in this study were more likely to
be satisfied with their jobs and were less likely to be embedded in their
jobs and to perceive alternative employment opportunities. We also found
that study participants who completed the survey at Time 2 were less likely
to perform internal networking behaviors and to perceive alternative em-
ployment opportunities than their nonresponsive counterparts. In general,
attrition bias threatens the validity of research findings by underestimat-
ing, overestimating, or generating spurious relationships between study
variables (Goodman & Blum, 1996). Given the prior empirical evidence
for many of the relationships addressed in this study, it is unlikely that
our findings are spurious; rather, the direct and indirect relationships of
networking and turnover antecedents with turnover may be either inflated
or understated, which limits the generalizability of these findings to all
individuals who network and leave their organizations. As such, future
research should replicate these findings in another sample.
26 PERSONNEL PSYCHOLOGY
Third, although we made an effort to include alternative explanatory
variables for turnover in our theorizing and as statistical controls, there
may still be factors that influence employees’ decisions to turnover (e.g.,
marital status, school-aged children, or financial inducements) for which
we were unable to account in our analyses. Thus, there is the possibil-
ity that there are alternative explanations for our findings. Finally, we
selected a sample working in a particular profession (i.e., industrial and
organizational psychologists) in order to achieve sufficient variance on the
networking measures, which may limit the generalizability of our findings
to other occupational contexts.
In addition to accounting for these limitations, future research should
build upon these findings to extend researchers’ knowledge of network-
ing behaviors and voluntary turnover phenomena. First, given that both
networking and turnover processes take place over long periods of time,
the relationships among networking and turnover antecedents are likely
more complicated than what we initially proposed and investigated in this
study. Future research should rigorously investigate dynamic interplay of
networking behaviors, work attitudes, and labor market perceptions as
they fluctuate over time to promote or deter voluntary turnover by adopt-
ing longitudinal research designs where networking behaviors, desires to
remain, and opportunities to exit are measured repeatedly (at least three
times). Such research designs would be capable of teasing apart whether
networking behaviors are associated with an increase in desires to remain
and opportunities to exit over time (cf. Wolff & Moser, 2009).
Second, although our study focuses on clarifying the conditions under
which networking behaviors may have potentially negative implications
for organizations (i.e., voluntary turnover) relative to the benefits that
accrue for individuals, future research should consider other potentially
dysfunctional consequences of networking behaviors to more fully un-
derstand the implications for organizations. As an example, under the
guise of networking, employees may engage in counterproductive work
behaviors (e.g., gossip) or other work-avoidant behaviors, which may
hinder departmental or organizational functioning. Related to this, future
research should systematically investigate whether the benefits of net-
working (e.g., increased exchange of instrumental information) outweigh
unintended outcomes (e.g., voluntary turnover). For example, research has
indicated that networking may function to isolate certain minority groups
(i.e., ethnic minorities, women), resulting in less access to resources than
their (White, male) counterparts (Ibarra, 1995). In this case, the benefits
of networking for the individual may have negative implications not only
for organizations but also for society.
Future research should also consider investigating the relative
advantages external networking poses for organizational functioning.
PORTER ET AL. 27
Our findings suggest that employees’ external networking behaviors
represent a potential risk to organizations because they are positively
associated with voluntary turnover. However, many benefits may result
from building, maintaining, and using relationships with professional
colleagues outside of one’s employing organization. For example,
within the current sample of industrial and organizational psychologists,
external networking may be used to exchange information related to
scientific advancements and/or benchmarking, which may translate
into an organizational advantage. Thus, future research may delve into
the specific types of interpersonal resources exchanged in networking
interactions to provide insights into whether and how employees’ external
networking behaviors may have positive implications for organizations.
Third, although bivariate correlations found in our study support the
suggestion that job satisfaction and perceived employment opportunities
are theoretically meaningful mediators in the networking–turnover rela-
tionships, they were no longer significant predictors of turnover when
considered simultaneously with other study variables in the structural
equation model. This is consistent with prior research findings that the
predictive relationship between job satisfaction and turnover disappears
when job embeddedness is included (Jiang et al., 2012) and that perceived
ease of movement no longer predicts turnover once actual obtainment
of alternative offer is also taken into account (Griffeth et al., 2005). Re-
search should further probe the relative importance of these constructs in
comparison with other attitudinal constructs (e.g., organizational commit-
ment, perceived organizational support) and economic factors (e.g., labor
market conditions) in predicting withdrawal and turnover to clarify which
variables have the most impact on one’s decisions to leave. In addition, it is
possible that perceived employment opportunities predict leaving without
a job offer in hand when employees have a strong desire to leave. There-
fore, there may be opportunities for future research to clarify whether
and how desires and alternative opportunities interact with one another to
predict different forms of voluntary turnover (e.g., Lee & Mitchell, 1994).
References
Adler PS, Kwon SW. (2002). Social capital: Prospects for a new concept. Academy of
Management Review,27, 17–40. doi: 10.2307/4134367
Ahearn K, Ferris GR, Hochwarter WA, Douglas C, Ammeter AP. (2004). Leader po-
litical skill and team performance. Journal of Management,30, 309–327. doi:
10.1016/j.jm.2003.01.004
Allen DG, Bryant PC, Vardaman JM. (2010). Retaining talent: Replacing misconceptions
with evidence-based strategies. Academy of Management Perspectives,May, 48–64.
doi: 10.5465/AMP.2010.51827775
Arthur MB, Rousseau DM (Eds.). (1996). The boundaryless career: A new employment
principle for a new organizational era. New York, NY: Oxford University Press.
28 PERSONNEL PSYCHOLOGY
Batt R, Colvin AJS. (2011). An employment systems approach to turnover: Human re-
sources practices, quits, dismissals, and performance. Academy of Management
Journal,54, 695–717. doi: 10.5465/AMJ.2011.64869448
Bian Y. (1997). Bringing strong ties back in: Indirect ties, network bridges, and job searches
in China. American Sociological Review,62, 366–385. doi: 10.2307/2657311
Blau G. (1993). Further exploring the relationship between job search and voluntary in-
dividual turnover. PERSONNEL PSYCHOLOGY,46, 313–330. doi: 10.1111/j.1744-
6570.1993.tb00876.x
Blau PM. (1964). Exchange and power in social life. New York, NY: John Wiley.
Borg IN, Groenen PJF. (2005). Modern multidimensional scaling. New York, NY: Springer.
Boswell WR, Boudreau JW, Tichy J. (2005). The relationship between employee job
change and job satisfaction: The honeymoon-hangover effect. Journal of Applied
Psychology,90, 882–892. doi: 10.1037//0021-9010.90.5.882
Boswell WR, Olson-Buchanan JB, Harris TB. (2014). I cannot afford to have a life:
Employee adaptation to feelings of job insecurity. PERSONNEL PSYCHOLOGY,67,
887–915. doi: 10.1111/peps.12061
Boswell WR, Zimmerman RD, Swider BW. (2012). Employee job search: Toward an
understanding of search context and search objectives. Journal of Management,38,
129–163. doi: 10.1177/0149206311421829
Brown TA. (2006). Confirmatory factor analysis for applied research. New York, NY:
Guilford Press.
Buhrmester M, Kwang T, Gosling SD. (2011). Amazon’s mechanical turk: A new source
of inexpensive, yet high-quality data? Perspectives on Psychological Science,6(1),
3–5. doi: 10.1177/1745691610393980
Burt RS. (1992). Structural Holes. Cambridge, MA: Harvard University Press
Cable DM, DeRue DS. (2002). The convergent and discriminant validity of subjective
fit perceptions. Journal of Applied Psychology,87, 875–884. doi: 10.1037//0021-
9010.87.5.875
Cable DM, Parsons CK. (2001). Socialization tactics and person-organization fit. PERSON-
NEL PSYCHOLOGY,54(1), 1–23. doi: 10.1111/j.1744-6570.2001.tb00083.x
Carlson DS, Kacmar KM. (2000). Work-family conflict in the organization: Do life
role values make a difference? Journal of Management,26, 1031–1054. doi:
10.1177/014920630002600502
Cascio WF. (2006). Managing human resources: Productivity, quality of work life, profits
(7th ed.). Burr Ridge, IL: Irwin/McGraw-Hill.
Chiaburu DS, Harrison DA. (2008). Do peers make the place? Conceptual synthesis and
meta-analysis of coworker effects on perceptions, attitudes, OCBs, and performance.
Journal of Applied Psychology,93, 1082–1103. doi: 10.1037/0021-9010.93.5.1082
Coleman JS. (1988). Social capital in the creation of human capital. American Journal of
Sociology,94, S94–S120. doi: 10.1086/228943
Crossley CD, Bennett RJ, Jex SM, Burnfield JL. (2007). Development of a global measure
of job embeddedness and integration into a traditional model of voluntary turnover.
Journal of Applied Psychology,92, 1031–1042. doi: 10.1037/0021-9010.92.4.1031
Davison ML, Skay CL. (1991). Multidimensional scaling and factor models of test
and item responses. Psychological Bulletin,110, 551–556. doi: 10.1037/0033-
2909.110.3.551
DiStefano C, Zhu M, Mˆ
ındrilˇ
a D. (2009). Understanding and using factor scores: Consid-
erations for the applied researcher. Practical Assessment, Research & Evaluation,
14(20), 1–11. Available at: http://pareonline.net/getvn.asp?v=14&n=20
Eby LT, Butts M, Lockwood A. (2003). Predictors of success in the era of the boundaryless
career. Journal of Organizational Behavior,24, 689–708. doi: 10.1002/job.214
PORTER ET AL. 29
Feldman JM, Lynch Jr JG. (1988). Self-generated validity and other effects of measurement
on belief, attitude, intention, and behavior. Journal of Applied Psychology,73, 421–
435. doi: 10.1037//0021-9010.73.3.421
Felps W, Mitchell TR, Hekman DR, Lee TW, Holtom BC, Harman WS. (2009).
Turnover contagion: How coworkers’ job embeddedness and job search behav-
iors influence quitting. Academy of Management Journal,52, 545–561. doi:
10.5465/AMJ.2009.41331075
Fernandez RM, Weinberg N. (1997). Sifting and sorting: Personal contacts and hiring in a
retail bank. American Sociological Review,62, 883–902. doi: 10.2307/2657345
Ferris GR, Treadway DC, Kolodinsky RW, Hochwarter WA, Kacmar CJ, Douglas C, Frink
DD. (2005). Development and validation of the political skill inventory. Journal of
Management,31(1), 126–152. doi: 10.1177/01149206304271386
Ferris GR, Witt LA, Hochwarter WA. (2001). Interaction of social skill and general mental
ability on job performance and salary. Journal of Applied Psychology,86, 1075–
1082. doi: 10.1037//021-9010.86.6.1075
Fisher CD. (1986). Organizational socialization: An integrative review. In KM Rowland,
GR Ferris (Eds.), Research in personnel and human resources management (Vol. 4,
pp. 101–145). Greenwich, CT: JAI Press.
Foa UG. (1971). Interpersonal and economic resources. Science,171, 345–351. doi:
10.1126/science.171.3969.345
Forret ML. (2013). Networking as a job search and career management behavior. In UC
Klehe, EWJ van Hooft (Eds.), The Oxford handbook of job loss and job search.New
York, NY: Oxford University Press.
Forret ML, Dougherty TW. (2001). Correlates of networking behavior and managerial
professional employees. Group & Organization Management,26, 283–311. doi:
10.1177/1059601101263004
Forret ML, Dougherty TW. (2004). Networking behaviors and career outcomes: Differ-
ences for men and women? Journal of Organizational Behavior,25, 419–437. doi:
10.1002/job.253
Goodman JS, Blum TC. (1996). Assessing the non-random sampling effects of sub-
ject attrition in longitudinal research. Journal of Management,22, 627–652. doi:
10.1177/014920639602200405
Gouldner AW. (1960). The norm of reciprocity: A preliminary statement. American Soci-
ological Review,25, 161–178. doi: 10.2307/2092623
Granovetter MS. (1995). Getting a job: A study of contacts and careers. Chicago, IL:
University of Chicago Press.
Grice JW. (2001). Computing and evaluating factor scores. Psychological Methods,6,
430–450. doi: 10.1037/1082-989X.6.4.430
Griffeth RW, Hom PW, Gaertner S. (2000). A meta-analysis of antecedents and
correlates of employee turnover: Update, moderator tests, and research impli-
cations for the next millennium. Journal of Management,26, 463–488. doi:
10.1177/014920630002600305
Griffeth RW, Steel RP, Allen DG, Bryan N. (2005). The development of a multidimen-
sional measure of job market cognitions: The employment opportunity index (EOI).
Journal of Applied Psychology 90, 335–349. doi: 10.1037/0021-9010.90.2.335
Harris KJ, Wheeler AR, Kacmar KM. (2011). The mediating role of organizational job
embeddedness in the LMX-outcomes relationships. The Leadership Quarterly,22,
271–281. doi: 10.1016/j.leaqua.2011.02.003
Hinkin TR. (1998). A brief tutorial on the development of measures for use in
survey questionnaires. Organizational Research Methods,1(1), 104–121. doi:
10.1177/109442819800100106
30 PERSONNEL PSYCHOLOGY
Holtom BC, Mitchell TR, Lee TW. (2006). Increasing human and social capital by ap-
plying job embeddedness theory. Organizational Dynamics,35, 316–331. doi:
10.1016/j.orgdyn.2006.08.007
Holtom BC, Mitchell TR, Lee TW, Eberly MB. (2008). Turnover and retention research: A
glance at the past, a closer review of the present, and venture into the future. Academy
of Management Annals,2(1), 231–274. doi: 10.1080/19416520802211552
Hom PW, Caranikas-Walker F, Prussia GE, Griffeth RW. (1992). A meta-analytical struc-
tural equations analysis of a model of employee turnover. Journal of Applied Psy-
chology,77, 890–909. doi: 10.1037/0021-9010.77.6.890
Hom PW, Mitchell T, Lee TW, Griffeth RW. (2012). Reviewing employee turnover: Focus-
ing on proximal withdrawal states and an expanded criterion. Psychological Bulletin,
138, 831–858. doi: 10.1037/a0027983
Hu LT, Bentler PM. (1999). Cutoff criteria for fit indexes in covariance structure analysis:
Conventional criteria versus new alternatives. Structural Equation Modeling,6(1),
1–55. doi: 10.1080/10705519909540118
Hulin C. (1991). Adaptation, persistence, and commitment in organizations. In MD Dun-
nette, LM Hough (Eds.), Handbook of industrial and organizational psychology
(Vol. 2, 2nd ed., pp.445–505). Palo Alto, CA: Consulting Psychologists Press.
Humphrey SE, Nahrgang JD, Morgeson FP. (2007). Integrating motivational, social, and
contextual work design features: A meta-analytic summary and theoretical extension
of the work design literature. Journal of Applied Psychology,92, 1332–1356. doi:
10.1037/0021-9010.92.5.1332
Ibarra H. (1995). Race, opportunity, and diversity of social circles in managerial networks.
Academy of Management Journal,38, 673–703. doi: 10.2307/256742
Jiang K, Liu D, McKay PF, Lee TW, Mitchell TR. (2012). When and how is job embed-
dedness predictive of turnover? A meta-analytic investigation. Journal of Applied
Psychology,97, 1077–1096. doi: 10.1037/a0028610
John OP, Srivastava S. (1999). The Big Five Trait taxonomy: History, measurement, and
theoretical perspectives. In LA Pervin, OP John (Eds.), Handbook of personality:
Theory and research (2nd ed, pp. 102–139). New York, NY: Guilford Press.
Judge TA, Hulin CL, Dalal RS. (2011). Job satisfaction and job affect. In SWJ Kozlowski
(Ed.), The Oxford handbook of industrial and organizational psychology.NewYork,
NY: Oxford University Press. doi: 10.1093/oxfordhb/9780199928309.013.0015
Kirmeyer SL, Lin TR. (1987). Social support: Its relationship to observed communication
with peers and superiors. Academy of Management Journal,30(1), 138–151. doi:
10.2307/255900
Lee TH, Gerhart B, Weller I, Trevor CO. (2008). Understanding voluntary turnover: Path-
specific job satisfaction effects and the importance of unsolicited job offers. Academy
of Management Journal,51, 651–671. doi: 10.5465/AMR.2008.33665124
Lee TW, Burch TC, Mitchell TR. (2014). The story of why we stay: A review of job
embeddedness. Annual Review of Organizational Psychology and Organizational
Behavior,1(3), 1–18. doi: 10.1146/annurev-orgpsych-031413-091244
Lee TW, Mitchell TR. (1994). An alternative approach: The unfolding model of vol-
untary employee turnover. Academy of Management Review,19, 51–89. doi:
10.2307/258835
Lee TW, Mitchell TR, Holtom BC, McDaniel LS, Hill JW. (1999). The unfolding model of
voluntary turnover: A replication and extension. Academy of Management Journal,
42, 450–462. doi: 10.2307/257015
Lin N. (2002). Social capital: A theory of social structure and action. Cambridge, MA:
Cambridge University Press. doi: 10.1017/CBO9780511815447
PORTER ET AL. 31
MacKinnon DP, Lockwood CM, Hoffman JM, West SG, Sheets V. (2002). A comparison
of methods to test mediation and other intervening variable effects. Psychological
Methods,7, 83–104. doi: 10.1037/1082-989X.7.1.83
MacKinnon DP, Lockwood CM, Williams J. (2004). Confidence limits for the indirect
effect: Distribution of the product and resampling methods. Multivariate Behavioral
Research,39(1), 99–128. doi: 10.1207/s15327906mbr3901_4
Maertz Jr CP, Campion MA. (2004). Profiles in quitting: Integrating process and
content turnover theory. Academy of Management Journal,47, 566–582. doi:
10.2307/20159602
Maertz Jr CP, Griffeth RW. (2004). Eight motivational forces and voluntary turnover: A
theoretical synthesis with implications for research. Journal of Management,30,
667–683.
March JG, Simon HA. (1958). Organizations. New York, NY: Wiley.
McCallum SY, Forret ML, Wolff H-G. (2014). Internal and external networking behav-
ior: An investigation of relationships with affective, continuance, and normative
commitment. Career Development International,19, 595–614. DOI: 10.1108CDI-
08-2013-0101
Mitchell TR, Holtom BC, Lee TW, Sablynski CJ, Erez M. (2001). Why people stay: Using
job embeddedness to predict voluntary turnover. Academy of Management Journal,
44, 1102–1121. doi: 10.2307/3069391
Mitchell TR, Lee TW. (2001). The unfolding model of voluntary turnover and job em-
beddedness: Foundations for a comprehensive theory of attachment. Research in
Organizational Behavior,23, 189–246. doi: 10.1016/S0191-3085(01)23006-8
Mobley WH, Griffeth RW, Hand HH, Meglino BM. (1979). Review and conceptual anal-
ysis of the employee turnover process. Psychological Bulletin,86, 493–522. doi:
10.1037/0033-2909.86.3.493
Morrison EW. (2002). Newcomers’ relationships: The role of social networks ties
during socialization. Academy of Management Journal,45, 1149–1160. doi:
10.2307/3069430
Mossholder KW, Settoon RP, Henagan SC. (2005). A relational perspective on turnover:
Examining structural, attitudinal, and behavioral predictors. Academy of Manage-
ment Journal,48, 607–618. doi: 10.5465/AMJ.2005.17843941
Muth´
en, LK. (2005, August). Interpreting results of model with dichotomous out-
come. Message posted to http://www.statmodel.com/discussion/messages/23/138.
html?1407284845
Nahapiet J, Ghoshal S. (1998). Social capital, intellectual capital, and the organizational
advantage. Academy of Management Review,23, 242–266. doi: 10.2307/259373
Ng TWH, Feldman DC. (2010). The effects of organizational embeddedness on develop-
ment of social and human capital. Journal of Applied Psychology,95, 696–712. doi:
10.1037/a0019150
Ployhart RE, Ward A-K. (2011). The “quick start guide” for conducting and publish-
ing longitudinal research. Journal of Business and Psychology,26, 413–422. doi:
10.1007/s10869-011-9209-6
Podsakoff PM, MacKenzie SB, Lee J-Y, Podsakoff NP. (2003). Common method biases in
behavioral research: A critical review of the literature and recommended remedies.
Journal of Applied Psychology,88, 879–903. doi: 10.1037/0021-9010.88.5.879
Porter CM, Woo SE. (2015). Untangling the networking phenomenon: A dynamic psycho-
logical perspective on how and why people network. Journal of Management,41,
1477–1500. doi: 10.1177/0149206315582247
Preacher KJ, Hayes AF. (2008). Asymptotic and resampling strategies for assessing and
comparing indirect effects in multiple mediator models. Behavior ResearchMethods,
40, 879–891. doi: 10.3758/BRM.40.3.879
32 PERSONNEL PSYCHOLOGY
Price JL, Mueller CW. (1981). A causal model of turnover for nurses. Academy of Man-
agement Journal,24, 543–565. doi: 10.2307/255574
Rogelberg SG, Stanton JM. (2007). Understanding and dealing with organizational
survey nonresponse. Organizational Research Methods,10, 195–209. doi:
10.1177/1094428106294693
Russell SS, Spitzmuller C, Lin LF, Stanton JM, Smith PC, Ironson GH. (2004). Shorter can
also be better: The abridged job in general scale. Educational and Psychological
Measurement,64, 878–893. doi: 10.1177/0013164404264841
Scott J. (2000). Social network analysis. Thousand Oaks, CA: Sage.
Steel RP. (2002). Turnover theory at the empirical interface: Problems of fit and function.
Academy of Management Review,27, 346–360. doi: 10.2307/4134383
Thompson JA. (2005). Proactive personality and job performance: A social capital
perspective. Journal of Applied Psychology,90, 1011–1017. doi: 10.1037/0021-
9010.90.5.1011
Tichy NM, Tushman ML, Fombrun C. (1979). Social network analysis for organizations.
Academy of Management Review,4, 507–519. doi: 10.2307/257851
Treadway DC, Breland JW, Adams GL, Duke AB, Williams LA. (2010). The interactive
effects of political skill and future time perspective on career and community net-
working behavior. Social Networks,32, 138–147. doi: 10.1016/j.socnet.2009.09.004
Van Hoye G, Lievens F. (2009). Tapping the grapevine: A closer look at word-of-
mouth as a recruitment source. Journal of Applied Psychology,94, 341–352. doi:
10.1037/a0014066
Van Hoye G, Van Hooft EAJ, Lievens F. (2009). Networking as a job search behaviour: a so-
cial network perspective. Journal of Occupational and Organizational Psychology,
82, 661–682. doi: 10.1348/096317908×360675
Velicer WF. (1976). Determining the number of components from the matrix of partial
correlations. Psychometrika,41, 321–327. doi: 10.1007/BF02293557
Vinokur A, Schul Y, Caplan RD. (1987). Determinants of perceived social support: Inter-
personal transactions, personal outlook, and transient affective states. Journal of Per-
sonality and Social Psychology,53, 1137–1145. doi: 10.1037//0022-3514.53.6.1137
Wanberg CR, Kanfer R, Banas JT. (2000). Predictors and outcomes of networking intensity
among unemployed job seekers. Journal of Applied Psychology,85, 491–503. doi:
10.1037//0021-9010.85.4.491
Wasserman S, Faust K. (1994). Social network analysis: Methods and applications.New
York, NY: Cambridge University Press.
Weiss HM, Cropanzano R. (1996). Affective events theory: A theoretical discussion of the
structure, causes and consequences of affective experiences at work. Research in
Organizational Behavior,18, 1–74. doi: 10.4135/9781452276090.n9
Williams LJ, Cote JA, Buckley MR. (1989). Lack of method variance in self-reported
affect and perceptions at work: Reality or artifact? Journal of Applied Psychology,
74, 462–468. doi: 10.1037/0021-9010.74.3.462
Wolff H-G, Kim S. (2012). The relationship between networking behaviors and the big
five personality dimensions. Career Development International,17(1), 43–66. doi:
10.1108/13620431211201328
Wolff H, Moser K. (2006). Development and validation of a networking scale. Diagnostica,
52, 161–180.
Wolff H, Moser K. (2009). Effects of networking on career success: a longitudinal study.
Journal of Applied Psychology,94(1), 196–206. doi: 10.1037/a0013350
Wolff H, Moser K. (2010). Do specific types of networking predict specific mobility
outcomes? A two-year prospective study. Journal of Vocational Behavior,77, 238–
245. doi: 10.1016/j.jvb.2010.03.001
PORTER ET AL. 33
Wolff H, Schneider-Rahm CI, Forret ML. (2011). Adaptation of a German multidimensional
networking scale into English. European Journal of Psychological Assessment,27,
244–250. doi: 10.1027/1015-5759/a000070
Woo SE, Maertz CP. (2012). Assessment of voluntary turnover in organizations: Answer-
ing the questions of why, who, and how much. In N Schmitt (Ed.), The Oxford
handbook of personnel assessment and selection (pp. 570–594). New York, NY:
Oxford University Press.
APPENDIX A
Shortening Networking Behavior Measures
Overview. The purpose of this study was to shorten networking be-
haviors measures to a manageable length. To ensure the reduced measures
adequately tap into the networking behaviors constructs, we examined
their relations to theoretically relevant personality constructs.
Participants and procedures. Data were collected through Amazon Me-
chanical Turk, which has been found to be a source of high-quality data
(e.g., Buhrmester, Kwang, & Gosling, 2011). We recruited 370 individuals
to participate in an online survey for a nominal reward. Participants’ aver-
age age was 36. Sixty-five percent of participants were women. Seventy-
seven percent of participants were White, 8% were Asian, 8% were Black,
and 4% were Hispanic. Thirty-four percent of respondents earned a bach-
elor’s degree, 30% attended some college or earned an associate’s degree,
18% had an advanced degree (master’s, doctoral, professional), and about
10% earned a high school diploma.
Measures. Information about the original networking behaviors scale
is reported in the main text of this manuscript. Extraversion, agreeable-
ness, conscientiousness, neuroticism, and openness to experience were
measured using 44 items from the Big Five Inventory (John & Srivastava,
1999) rated on a five-point Likert scale ranging from 1 =strongly disagree
to 5 =strongly agree.
Item reduction. To identify items for removal, we ran an exploratory
factor analysis (maximum likelihood estimation with an oblique, promax
rotation) on Wolff and Moser’s (2006) original internal and external net-
working behavior items. Using Velicer’s MAP method of parallel analysis,
five factors were retained (Velicer, 1976). However, this factor solution
was not easily interpretable as the items did not cleanly load on factors
representing the six networking behaviors types produced by the original
scale (Wolff & Moser, 2006). There are two complementary explanations
for these findings. First, this factor solution revealed that certain internal
and external networking behaviors items that tap into the same network-
ing function (e.g., maintaining) loaded on the same factor, which is due
to the fact that network relationship development, regardless of the focus
34 PERSONNEL PSYCHOLOGY
of networking being with internal or external contacts, follows the same
underlying process of building, maintaining, and using network contacts
(Porter & Woo, 2015). That is, the actual behaviors that individuals use to
build, maintain, and use relationships with internal and external contacts
are quite similar. Given the similarities in the behaviors themselves, it is
difficult to empirically detect fine grained distinctions between the foci of
networking behaviors (i.e., internal or external contacts) by investigating
the behaviors in isolation (e.g., using exploratory factor analysis).
Second, these findings may also be attributed to the fact that this
set of scales was originally developed through multidimensional scaling
(MDS; Wolff et al., 2011), which has different underlying assumptions
and purposes than traditional psychometric techniques based on the com-
mon factor model. Whereas traditional EFA is used to identify a set of
latent factors that are clearly distinguishable from one another based on
their shared variance, MDS is used to identify underlying dimensions that
explain similarities (or distances) between a set of variables in geometric
space (Borg & Groenen, 2005). Using MDS, the original scale was found
to have three dimensions, one representing item difficulty, a second repre-
senting the structural facet of internal and external networking behaviors,
and a third representing the functional facets of building, maintaining, and
using. Thus, these items are likely to be strongly related to one another,
and therefore, difficult to empirically distinguish using EFA, as is often
the case when a scale is developed using MDS (Davison & Skay, 1991).
Following recommendations by Hinkin (1998), we evaluated multi-
ple criteria to identify items that were suitable candidates for removal.
Namely, we considered (1) cross-loadings (i.e., whether an item loaded
on multiple factors with similar magnitudes); (2) interitem correlations
(i.e., whether items showed strong correlations with other items: interitem
correlations for internal networking behaviors ranged from .22 to .69,
and those for external networking behaviors ranged from .16 to .62); (3)
initial item communalities (i.e., whether items showed strong commu-
nalities, which ranged from .54 to .64 for internal networking behaviors
and from .52 to .70 for external networking behaviors); and (4) whether
items adequately sampled from the content domain (i.e., all functional
dimensions). Based on these criteria, we retained nine items for each
scale (example items are reported in Table 1). We then conducted another
round of exploratory factor analysis on the retained items. Velicer’s MAP
method of parallel analysis suggested that a one-factor solution was most
appropriate. We attributed this finding to the similarities in the networking
behaviors and the original scale development technique.
Confirmatory factor analysis. We also conducted a confirmatory factor
analysis of the selected items (using the same dataset described above)
to evaluate whether a six-factor model (with separate structural and
PORTER ET AL. 35
TABL E A1
Means, Standard Deviations, Reliabilities, and Intercorrelations of Study Variables
MSD 123 456789
1. Age 36 12.50 –
2. Gender .35 .48 −.02 –
3. Extraversion 3.10 .88 .03 .00 (.91)
4. Neuroticism 2.60 .80 −.17∗∗ −.04 −.37∗∗ (.89)
5. Conscientiousness 4.01 .64 .21∗∗ −.12∗.28∗∗ −.50∗∗ (.85)
6. Agreeableness 3.86 .59 .16∗∗ −.08 .27∗∗ −.51∗∗ .41∗∗ (.80)
7. Openness 3.78 .60 .11∗.02 .21∗∗ −.10 .18∗∗ .19∗∗ (.84)
8. Internal networking 2.28 .61 .03 .13∗.51∗∗ −.21∗∗ .14∗∗ .24∗∗ .32∗∗ (.87)
9. External networking 2.05 .64 −.02 .13∗.41∗∗ −.17∗∗ .05 .15∗∗ .25∗∗ .76∗∗ (.88)
Note. Gender: 0 =female, 1 =male. Reliability estimates are along the diagonal.
∗p<.05. ∗∗p<.01.
36 PERSONNEL PSYCHOLOGY
functional facets), a three-factor model (with separate functional facets),
a two-factor model (with separate structural facets), or a one-factor model
best fit the data. We found that the six- and three-factor models converged
to inadmissible solutions; the correlation matrices were not positive def-
inite. When we compared the other two models, we found that the two-
factor model (χ2(134, N=370) =507.32, p<.001, CFI =.88, RMSEA =
.09) fit better than the one-factor model (χ2(1, N=370) =98.48, p<
.01; χ2(135, N=370) =605.80, p<.001, CFI =.85, RMSEA =.10).
Using the data set from the main study, we found similar results: The six-
and three-factor models were not tenable due to a nonpositive definite
matrix, and we found that the two-factor (χ2(134, N=349) =559.18,
p<.001, CFI =.86, RMSEA =.10) model provided a better fit to the data
than a one-factor model (χ2(1, N=349) =880.92, p<.01; χ2(135,
N=349) =1440.10, p<.001, CFI =.67, RMSEA =.15). Although
the fit indices do not reach commonly accepted cut-off values (e.g., Hu
& Bentler, 1999), these findings suggest that the refined networking be-
haviors scales are capable of distinguishing between internal and external
networking behaviors.
Relationships with personality variables. To further evaluate the con-
struct validity of the reduced scales, we examined bivariate relationships
between internal and external networking behaviors and the Big Five
personality characteristics. Means, standard deviations, scale reliabilities,
and intercorrelations of study variables are shown in Table A1. Gener-
ally, the pattern of results was consistent with previous research (Forret
& Dougherty, 2001; Wolff & Kim, 2012): Both internal and external
networking were positively related to extraversion, agreeableness, and
openness to experience and negatively related to neuroticism. From a the-
oretical standpoint, we expect one’s personality characteristics to relate
to both internal and external networking similarly as the underlying pro-
cess of networking (i.e., building and maintaining network contacts) is
the same regardless of whether one is networking with contacts within or
outside of one’s employing organization (Porter & Woo, 2015). As such,
when considered separately, internal networking or external networking
likely captures an individual’s tendency to network, in general.
APPENDIX B
A Modified Two-Factor Measurement Model for Internal and External
Networking Behaviors
Given that the theoretical emphasis of the main study is on teasing apart
the unique relationships of internal and external networking with turnover
PORTER ET AL. 37
antecedents and turnover, we adopted a modified two-factor measurement
model of networking where internal and external networking are specified
as two separate factors (each represented by building, maintaining, and
using scale scores) as an alternative to the six-factor measurement model
that was originally proposed by Wolff and Moser (2006). For the cur-
rent research purposes, distinguishing between the specific functions of
networking was of less concern, because the same underlying process of
networking (represented by building, maintaining, and using functions) is
applicable regardless of whether one is networking with internal or exter-
nal network contacts. Rather, our emphasis was on with whom employees
were networking. Networking directed toward a specific set of network
contacts (either internal or external) generates access to different types
of resources from each set of contacts, and it is through these different
resources that internal and external networking have their differential re-
lationships with employee outcomes (regardless of which function is at
play).
To empirically evaluate whether it was appropriate to use the refined
scales to assess the structural dimensions of internal and external network-
ing behaviors, we compared our hypothesized two-factor measurement
model where internal and external networking are modeled as separate
factors with building, maintaining, and using scale scores as indicators
(χ2(8, N=349) =36.29, p<.001; CFI =.97, TLI =.95, RMSEA =
.10, SRMR =.05) to a one-factor model (χ2(9, N=349) =284.40, p<
.001; CFI =.73, TLI =.55, RMSEA =.30, SRMR =.12). We found that
the two-factor model fit substantially better (χ2(1, N=349) =248.11,
p<.001), providing evidence that internal and external networking be-
haviors are distinct from one another in our data.
When we adopted this modified measurement model, there was evi-
dence that internal and external networking exhibited similar relationships
with the same construct when considered individually as well as moder-
ate to strong relationships with one another in both the scale refinement
(see Appendix A) and main study samples. Here we discuss why these
findings emerged, which we hope will provide insights for future research
investigating internal and external networking behaviors. First, we found
that internal and external networking behaviors had similar bivariate rela-
tionships with study outcomes, but they had differential relationships with
these outcomes when considered simultaneously, which is similar to the
findings of previous research investigating internal and external network-
ing (e.g., MacCallum, Forret, & Wolff, 2014). These findings are likely
a function of both internal and external networking measures assessing
a general behavioral pattern of networking. That is, when considered in-
dividually, the unique effects of internal or external networking may not
38 PERSONNEL PSYCHOLOGY
be apparent. Thus, to isolate the unique effects of internal and external
networking on individual outcomes, multivariate analyses are necessary.
Second, we also found moderate to strong relationships between inter-
nal and external networking behaviors in both the main study and the scale
refinement study samples. We attribute these findings to the fact that both
internal and external networking follow the same underlying networking
process (i.e., building, maintaining, and using contacts). However, we ar-
gue that the similarities between the underlying processes of internal and
external networking should not obscure the theoretical distinctiveness of
these two forms of networking: As mentioned above, internal and exter-
nal networking behaviors are conducted with two distinct sets of contacts
that offer access to different resources, which differentially contribute to
employee outcomes. Moreover, consistent with this study, in the applied
psychology and management literatures, there are multiple instances of
constructs that are strongly associated with one another but that differen-
tially predict criteria and are, therefore, considered theoretically distinct
(e.g., person–organization fit and needs–supplies fit; Cable & DeRue,
2002).
In general, we argue that both the functional (i.e., building, main-
taining, and using) and structural (i.e., internal and external) dimensions
of networking are valuable and valid approaches to conceptualizing and
investigating networking, and the choice of which conceptualization is
appropriate is dependent upon the theoretical emphasis of a study. For
instance, differentiating specific functions of networking behaviors is par-
ticularly useful when a study is geared toward understanding the processes
underlying network relationship development over time, as each function
plays an integral role in relationship initiation and growth (Porter & Woo,
2015). On the other hand, distinguishing between internal and external
networking is more appropriate for studies focused on delineating the
unique roles of different sets of network contacts in promoting or deter-
ring individual outcomes.
