Effects of Networking on Career Success: A Longitudinal Study

Abstract

Previous research has reported effects of networking, defined as building, maintaining, and using relationships, on career success. However, empirical studies have relied exclusively on concurrent or retrospective designs that rest upon strong assumptions about the causal direction of this relation and depict a static snapshot of the relation at a given point in time. This study provides a dynamic perspective on the effects of networking on career success and reports results of a longitudinal study. Networking was assessed with 6 subscales that resulted from combining measures of the facets of (a) internal versus external networking and (b) building versus maintaining versus using contacts. Objective (salary) and subjective (career satisfaction) measures of career success were obtained for 3 consecutive years. Multilevel analyses showed that networking is related to concurrent salary and that it is related to the growth rate of salary over time. Networking is also related to concurrent career satisfaction. As satisfaction remained stable over time, no effects of networking on the growth of career satisfaction were found. (PsycINFO Database Record (c) 2009 APA, all rights reserved).

Full text

Effects of Networking on Career Success: A Longitudinal Study
Hans-Georg Wolff
University of Erlangen-Nuremberg
Klaus Moser
University of Erlangen-Nuremberg
(December 2008)
LASER Discussion Papers - Paper No. 24
(edited by A. Abele-Brehm, R.T. Riphahn, K. Moser and C. Schnabel)
Correspondence to:
Dr. Hans-Georg Wolff, Lange Gasse 20, 90403 Nuremberg, Germany, Email:
hans-georg.wolff@wiso.uni-erlangen.de.

Abstract
Previous research has reported effects of networking, defined as building, maintaining, and using
relationships, on career success. However, empirical studies have exclusively relied upon concurrent
or retrospective designs that rest upon strong assumptions on the causal direction of this relation and
also depict a static snapshot of the relation at a given point in time. This study provides a dynamic
perspective of the effects of networking on career success and reports results of a longitudinal study.
Networking was assessed by six subscales that result from combining the facets of 1) internal vs.
external networking and 2) building vs. maintaining vs. using contacts. Objective (salary) and
subjective (career satisfaction) measures of career success were obtained for three consecutive years.
Multilevel analyses show that networking is related to concurrent salary and moreover, that it is
related to the growth rate of salary over time. Networking is also related to concurrent career
satisfaction. As satisfaction remained stable over time, no effects of networking on the growth of
career satisfaction were found.
Copyright statement
The copyright of this document has been transferred on behalf of the authors to the American
Psychological Association (APA), Journal of Applied Psychology. This article may not exactly
replicate the final version published in the APA journal. It is not the copy of record. This document
has been posted for the purpose of discussion and rapid dissemination of preliminary research results.
Author note
The authors would like to thank Jeff Johnson and James LeBreton for their advice on relative weights
analyses. Research reported in this paper has been supported by the Hans-Frisch-Foundation.

Networking and Career Success
1
© by American Psychological Association
http://www.apa.org/journals/apl
Running head: NETWORKING AND CAREER SUCCESS
Effects of Networking on Career Success: A Longitudinal Study
Draft Version June, 6th 2008
Copyright statement
The copyright of this document has been transferred on behalf of the authors to the American
Psychological Association (APA), Journal of Applied Psychology. This article may not
exactly replicate the final version published in the APA journal. It is not the copy of record.
This document has been posted for the purpose of discussion and rapid dissemination of
preliminary research results.
Author note. The authors would like to thank Jeff Johnson and James LeBreton for
their advice on relative weights analyses. Research reported in this paper has been supported
by the Hans-Frisch-Foundation.

Networking and Career Success
2
Abstract
Previous research has reported effects of networking, defined as building,
maintaining, and using relationships, on career success. However, empirical studies have
exclusively relied upon concurrent or retrospective designs that rest upon strong assumptions
on the causal direction of this relation and also depict a static snapshot of the relation at a
given point in time. This study provides a dynamic perspective of the effects of networking
on career success and reports results of a longitudinal study. Networking was assessed by six
subscales that result from combining the facets of 1) internal vs. external networking and 2)
building vs. maintaining vs. using contacts. Objective (salary) and subjective (career
satisfaction) measures of career success were obtained for three consecutive years. Multilevel
analyses show that networking is related to concurrent salary and moreover, that it is related
to the growth rate of salary over time. Networking is also related to concurrent career
satisfaction. As satisfaction remained stable over time, no effects of networking on the
growth of career satisfaction were found.
Key Words: Networking, career success, career development, social interaction

Networking and Career Success
3
Effects of networking on career success: A longitudinal study
Many books and articles in the practitioner literature suggest that networking
behaviors, such as going for out for drinks to discuss business matters informally, attending
conferences, or staying in contact with former colleagues, are essential to career success (e.g.,
Nierenberg, 2002; Torres, 2005; Welch, 1980). Similarly, scholarly research has shown that
networking is positively related to objective and subjective measures of career success (Forret
& Dougherty, 2004; Langford, 2000; Michael & Yukl, 1993; Orpen, 1996). Networking is
also associated with favorable performance ratings (Sturges, Conway, Guest, & Liefhooghe,
2005; Thompson, 2005) and may be a viable job search strategy (Wanberg, Kanfer, & Banas,
2000). Networking behaviors are used to build and maintain informal contacts that enhance
career success (Forret & Dougherty, 2004; Luthans, Rosenkrantz, & Hennessey, 1985;
Michael & Yukl, 1993).
To our knowledge, all studies on the relationship between networking and career
success have utilized either concurrent or retrospective designs that are not without
limitations. Concurrent designs provide a static snapshot of the relation between networking
and career success since they show that networkers are more successful than non-networkers
at a given point in time. These designs do not provide strong evidence for causality, i.e., that
networking has led to career success. Furthermore, concurrent designs do not yield insights
into the dynamics of this relation and have ignored effects of networking on the change of
career success over time. Popular theorizing typically assumes that networking is associated
with accelerated growth in career success, e.g., that the salary gap between networkers and
non-networkers increases over time. The examination of this dynamic effect requires the
observation of individual trajectories of career success over time and cannot be answered by
concurrent research designs. Retrospective designs provide the opportunity to study dynamic

Networking and Career Success
4
effects since they relate networking to prior career success. For example, Michael and Yukl
(1993) assessed the number of promotions an individual had received in his or her career.
However, these designs also rely on strong assumptions because they do not take the proper
temporal order of variables into account. They implicitly assume that networking leads to
career success, but cannot rule out the possibility that it is necessary to resort to networking
as one climbs up the career ladder and has to fulfill tasks of higher responsibility and
discretion (see e.g., Katz & Kahn, 1978).
The purpose of the present study is to overcome these limitations by investigating the
effects of networking on career success using a longitudinal research design. We therefore
take the presumed causal order, from networking to career success, into account. Moreover,
by examining individual trajectories of career success over time, we do not just examine
whether networking is related to career success, but also whether networking is related to
accelerated growth in career success. The study contributes to the literature in two aspects.
First, we further investigate the causal link between networking and career success and
provide stronger evidence for the causal influence of networking on career success. Second,
we emphasize the notions of time and change in our study and thus introduce a dynamic
perspective into networking research (e.g., Raudenbush, 2001).
Networking
The current research defines networking by behaviors that are aimed at building,
maintaining, and using informal relationships that possess the (potential) benefit to facilitate
work related activities of individuals by voluntarily granting access to resources and
maximizing common advantages (Wolff & Moser, 2006; see also Forret & Dougherty, 2004).
The construct is defined on a behavioral level (e.g., Michael & Yukl, 1993; Wanberg et al.,
2000; Witt, 2004) and can be considered a “behavior syndrome” (cf. Frese, Fay, Hilburger,
Leng, & Tag, 1997), that is, a set of interrelated behaviors that are consistently shown by

Networking and Career Success
5
individuals. Accordingly, networking measures typically assess how often individuals show
networking behaviors, e.g., discussing business matters outside of working hours or using
contacts to get confidential advice. Theoretical accounts (e.g., Cohen & Bradford, 1989;
Kaplan, 1984; Michael & Yukl, 1993) assume that these behaviors lead to informal,
voluntary, and reciprocal relationships that in turn facilitate access to resources such as task
related support, strategic information, or career success (Podolny & Baron, 1997; Wolff,
Moser, & Grau, in press).
Networking is distinct from the concept of social capital, which refers to a different
level of analysis. Networking is an individual level construct and focuses on individual
behavior. The concept of social capital refers to a structural level of analysis and focuses on
the quality and extent of existing relationship constellations (Adler & Kwon, 2002; Burt,
1992; Seibert, Kraimer, & Liden, 2001). For example, Coleman (1988) states that “social
capital inheres in the structure of relations between actors and among actors” (p. S98). It is
therefore closely linked to the position of an individual in a network and is typically
characterized by specific aspects of network structures such as network size, density, or
structural holes. In contrast, networking emphasizes individual actions and assesses to what
extent individuals proactively build and develop contacts. Networking can thus be considered
one out of several predictors of network structures (Wolff & Moser, 2006). However, social
capital also depends on situational opportunities to a high extent (Burt, 1992); for example,
holding a supervisory position (Carroll & Teo, 1996) or a position of high workflow
criticality (Brass, 1984).
Networking and Career success
Career success is defined as the “positive psychological or work-related outcomes or
achievements one has accumulated as a result of one’s work experiences” (Judge, Cable,
Boudreau, & Bretz, 1995 p. 486). In accordance with other studies (e.g., Gattiker & Larwood,

Networking and Career Success
6
1989; Judge et al., 1995; Seibert et al., 2001), we distinguish objective career success from
subjective success. Objective career success refers to observable career accomplishments that
can be reliably judged by others, such as pay and ascendancy. Subjective career success is
more concerned with individual appraisals of one’s career success. This subjective judgment
is not only influenced by objective criteria but also by individual aspiration levels, social
comparisons to relevant others, and situational constraints such as opportunities for
advancement in a profession (e.g., Arthur, Khapova, & Wilderom, 2005; Gattiker &
Larwood, 1989).
Several studies show that networking is related to both objective and subjective career
success (Forret & Dougherty, 2004; Langford, 2000; Michael & Yukl, 1993; Orpen, 1996).
For example, Michael and Yukl (1993) found that networking is related to the number of
promotions an individual has received in her or his career, and Langford (2000) showed that
networking is related to perceived career success. Unfortunately, these studies have all used
either concurrent or retrospective designs that suffer from the limitations described in the
introduction section. They only provide limited support for causal evidence and ignore the
dynamics of career success over time.
We therefore argue that a dynamic perspective on the relation between networking
and career success is necessary. Theories linking networking to career success assume that
networking is a way to get ahead (of others), which not only implies static differences in
career success but also accelerated growth of career success. We use a longitudinal design to
disentangle the concurrent (i.e., static) effects from growth (i.e., dynamic) effects of
networking over time. We assume that networking is related to concurrent career success,
thus replicating results from concurrent research designs. Furthermore, extending prior
research, networking should also be related to the prospective growth of career success. We
will use salary, the most prominent indicator of objective career success (see, e.g., the meta-

Networking and Career Success
7
analysis by Ng, Eby, Sorensen, & Feldman, 2005) as a measure of objective success. In
addition, we will also examine individuals’ career satisfaction to assess subjective career
success (Judge et al., 1995; Ng et al., 2005; Seibert et al., 2001). If networking leads to
objective career success as well as accelerated growth of success, this should also result in
increased satisfaction with one’s career. Career satisfaction may even be enhanced as
networking leads to a broad network of contacts with more opportunities to compare one’s
individual accomplishments with those of other individuals.
Hypothesis 1: Networking is related to concurrent objective career success.
Hypothesis 2: Networking is related to growth of objective career success.
Hypothesis 3: Networking is related to concurrent subjective career success.
Hypothesis 4: Networking is related to growth of subjective career success.
Method
Participants and Procedure
The study used a panel design with three survey waves and was conducted in
Germany. In October 2001, we collected addresses from 455 employed individuals that we
asked to participate in a longitudinal study on predictors of career success. To evade
problems of restricted sampling range, we used several means to recruit participants; e.g., we
were able to include invitations to participate in our study in official letters to participants of
(non-university based) night school training classes as well as to university alumni. We also
approached participants at official events, e.g., an alumni party, and asked personal contacts
to approach employees in their company. We mailed questionnaires to these 455 individuals,
of which 279 were returned, for a response rate of 61.3 %. Respondents’ mean age was 32
years (SD = 6.5), 60.4% were male, and 42% had a college degree. Participants came from a
wide range of industry sectors, the most frequent were the service industry (42%),
manufacturing (30%), and trade organizations (14%). Questionnaires were mailed to all

Networking and Career Success
8
addresses again in November 2002 and December 2003, where 227 (81.7% of 279
participants) and 202 (72.4%) questionnaires were returned, respectively. Following
Goodman and Blum (1996), we conducted dropout analyses by regressing dichotomous
indicators of missingness for Waves 2 and 3 on our study variables. Analyses showed no
systematic dropout at Wave 2. However, females and individuals from larger organizations
were significantly less likely to participate in the third wave. While this does indicate
systematic dropout at Wave 3, it is important to note that this dropout is only related to
control variables, but dropout does not depend on networking or career success (Menard,
1991).
Three substantive criteria were used to select participants for the analyses. First, we
only included predominantly working respondents who worked more than 20 hours per week
and earned more than €5000 (roughly US$ 5000 in the observation period) per year. Second,
we included only those respondents who were permanently employed during the observation
period, excluding participants for a variety of reasons (e.g., maternity leave, spells of
unemployment, or severe illness). While analyses for these participants would be of interest,
their small number and their highly specific situations rendered a substantive analysis
impossible (e.g., three women took maternity leave during the study period). Finally,
participants with missing values at Wave 1 in control variables, networking, or career success
variables were excluded.
We also included participants with partially missing data in the dependent variables at
Waves 2 and 3 in our analyses. One advantage of the multilevel analyses we used is that
participants with missing data in the dependent variable at some survey waves can be
included in the analyses (i.e., information from participants who provided data at Wave 1 and
Wave 2, but not Wave 3 can be included). This method provides better estimates of
regression coefficients than the usual listwise deletion method (Maas & Snijders, 2003;

Networking and Career Success
9
Schafer & Graham, 2002). Therefore, our sample size for multilevel analyses is N = 235,
which is higher than the number of subjects who responded at Wave 3 (i.e., N = 202).
Measures
Networking. Networking was measured with a German 44-item measure developed by
Wolff and Moser (2006).1 Similar to other networking measures (e.g., Forret & Dougherty,
2001; Michael & Yukl, 1993), this measure is multidimensional and is based upon two
theoretically derived facets: 1) a structural facet of internal vs. external networking and 2) a
functional facet of building vs. maintaining vs. using contacts. Crossing these facets leads to
six scales, building internal contacts (6 items, e.g., “I use company events to make new
contacts” α = .76), maintaining internal contacts (8 items, e.g., “I catch up with colleagues
from other departments about what they are working on” α = .69),2 using internal contacts (8
items, e.g., “I use my contacts with colleagues in other departments in order to get
confidential advice in business matters” α = .75), building external contacts (7 items, e.g., “I
accept invitations to official functions or festivities out of professional interest” α = .82),
maintaining external contacts (7 items, e.g., “I ask others to give my regards to business
acquaintances outside of our company” α = .76), and using external contacts (8 items, e.g., “I
exchange professional tips and hints with acquaintances from other organizations” α = .76).
All items were answered on a 4-point Likert scale ranging from 1 (never/very seldom) to 4
(very often/ always). In three studies, Wolff and Moser (2006) have provided evidence for the
validity of these scales and also for their differential validity, for example, they show that
generalized trust expectations (i.e., Interpersonal Trust) are more closely related to building
contacts than to maintaining or using contacts, where specific trust expectations concerning
specific relations become more important.
To provide further evidence on the construct validity of the scale, we conducted
several confirmatory factor analyses using item parcels due to the high number of items in

Networking and Career Success
10
relation to subjects. To avoid “data snooping” (Little, Cunningham, Shahar, & Widaman,
2002 p. 161) we followed the suggestion of Little et al. and used the same item parcels that
had been used by Wolff and Moser (2006) in a similar analysis. Results show satisfactory fit
for a correlated six factor model (Chi² (174) = 246.15; RMSEA = 0.040; CFI= 0.98). In
addition, this model provided better fit to the data than other models; for example, models
distinguishing either the structural (i.e., two factors: internal/ external networking; Chi² (188)
= 1005.65; RMSEA = 0.120; CFI = 0.79; two vs. six factor model: ΔChi² (14) = 759.5; p <
.01) or the functional facet (i.e., three factors: building/ maintaining/ using contacts; Chi²
(186) = 670.78; RMSEA = 0.096; CFI = 0.87; three vs. six factor model: ΔChi² (12) = 424.6;
p < .01), or a one factor model (Chi² (194) = 1225.17; RMSEA = 0.144; CFI = 0.70; one vs.
six factor model: ΔChi² (20) = 979.0; p < .01).
Objective career success. Following Judge et al. (1995), participants were asked to
report their gross yearly salary including bonuses, stock options, and other forms of cash
compensation. At Wave 1, participants reported their year 2000 salary in German Marks
(DM) that we use as a measure of concurrent salary. At Waves 2 and 3 participants specified
their annual salary either in DM or Euro (€), as Germany changed its currency from German
Marks (DM) to Euro (€) on Dec. 31st, 2001. All data were converted to Euro using the official
exchange rate of 1.95 DM to 1 €. In addition, participants were asked to provide information
on annual salaries for two years at Waves 2 and 3. At Wave 2, in 2002, we asked participants
to recall their salary from the previous year and also asked participants to estimate their
annual salary for the present year. Similarly, at Wave 3 we asked respondents to recall their
2002 salary and to estimate their 2003 salary. As the questionnaires were mailed close to the
end of the year, we assumed that participants could reliably estimate their annual salary for
the present year. We tested this assumption for the 2002 salary where two measures assessed
one year apart from each other were available: One salary estimate provided at the end of

Networking and Career Success
11
2002 (Wave 2), and one estimate recalled in 2003 (Wave 3). The correlation between these
two estimates was r = .96 with a small and insignificant mean difference of € 1532.48 (t
(116) = 0.52, p = .61; d = -0.06). As this indicates that participants can reliably estimate their
salary at the end of the year, we decided to use the estimate for 2003 as a fourth measure of
salary, i.e., salary estimates were available for 2000, 2001, 2002, and 2003. Due to deviations
from the normal distribution, we used the natural logarithm of salary in our analyses (see
Judge et al., 1995).
Subjective career success. We used the translation-backtranslation method to obtain a
German version of the career satisfaction scale by Greenhaus, Parasuraman, and Wormley
(1990) to measure subjective career success. The scale consists of five items (e.g., “I am
satisfied with the success I have achieved in my career”) and participants indicated their
agreement on a five-point scale ranging from 1 (do not agree at all) to 5 (fully agree).
Confirmatory factor analysis of the data from the first wave showed that a single factor model
fits the data well (χ² (5) = 9.17, p = .08; RMSEA = 0.059; CFI = 1.00). Career satisfaction
was assessed at each of the three waves. The reliability of this scale was α = .84 at each
survey wave.
Control variables. Several additional variables were included in the study to control
for factors that might confound the relationship between networking and career success
(Becker, 2005). We assessed two organizational variables, organizational size and whether
participants were in a supervisory position at the beginning of the study (0 = no, 1 = yes)
because these variables may influence opportunities to network (Forret & Dougherty, 2001)
as well as career success. In addition, three human capital variables, education, job
experience, and organizational tenure, were used. Education is related to network size
(Carroll & Teo, 1996) and also salary. Ng et al. (2005) have shown that experience and
tenure are related to salary and may also influence networking behavior (e.g., Kram &

Networking and Career Success
12
Isabella, 1985). Also, we controlled for two demographic variables that have been reported to
correlate with career success (Ng. et al., 2005), gender (1 = female, 2 = male), and
relationship status (1 = having a partner, 2 = single). Finally, in regressing career satisfaction
on the networking scales, we controlled for the natural logarithm of salary (e.g., Judge et al.,
1995). To preserve a meaningful intercept, we centered salary in this analysis. This did not
affect other regression coefficients.
Analyses
In the present analyses, measurement occasions are nested within participants. We
therefore used multilevel analysis (Bryk & Raudenbush, 2002; Snijders & Boskers, 1999).
This type of analysis allows for estimation of a trajectory of individual change in career
success and for differentiation between concurrent career success and the change of success
over time. As recommended for analyses with few measurement occasions, we used a linear
growth model (Bryk & Raudenbush, 2002) and coded time using consecutive integers
starting from zero.3 In this model, the intercept coefficient of the level 1 equation depicts
concurrent career success and the level 1 slope coefficient provides an estimate of linear
change over time, i.e., the growth of career success. In predicting career satisfaction, we
entered annual salary as a time varying level 1 control variable into the multilevel model. All
other control variables as well as the networking scales represent level 2 variables in our
model.
As shown in the measures section, we obtained a correlated factors solution for the
networking scales. Therefore, multicollinearity may pose a problem and the interpretation of
multilevel regression estimates in a “parameter-by-parameter fashion must proceed with
caution” (Shieh & Fouladi, 2003 p. 956). Partial redundancy of predictors may yield few
significant parameters, and in generalizing these results, sample to sample variation may lead
to different results with regard to the significance of particular predictors in other studies

Networking and Career Success
13
(Shieh & Fouladi, 2003). To avoid these problems, we based our interpretation upon two
procedures. First, we used a hierarchical approach in our analyses and examined the
difference in deviance (Δ deviance, see e.g., Bryk & Raudenbush, 2002) to determine if the
six networking scales improved model fit as a variable set (Cohen, Cohen, Aiken, & West,
2003).
Second, to examine the relative importance of each scale, we calculated relative
weights following a procedure by Johnson (2000; see also LeBreton, Harpis, Griepentrog,
Oswald, & Ployhart, 2007). This procedure is based on the calculation of a full principal
components solution from the original variables, where the components are rotated to an
orthogonal solution as similar as possible to the original variables (i.e., each variable has a
high loading on only one component). The orthogonal components resemble linear
transformations of the original networking scales; that is, together they carry the same
amount of information as the original variables. These components are then used as
predictors in the multilevel model – as they are orthogonal, multicollinearity is no concern in
this analysis. The relation between the orthogonal components and the original variables is
established by a regression of the original variables on the orthogonal components. Relative
weights are calculated by summing the product of the squared regression coefficients
between a) original variables and orthogonal components and b) orthogonal components and
the dependent variables from the multilevel model. 4 Relative weights are transformed into
proportions by dividing them by the sum of the total effects, which yields a proportional
contribution of each original variable (e.g., Johnson, 2000).
To further examine relative weights, we used a bootstrapping procedure with 1000
bootstrap samples to construct confidence intervals for the relative weights (Johnson, 2004).
As suggested by Johnson, we used the empirically derived confidence intervals from the
bootstrap (i.e., α/2 percentiles), because the distribution of the weights deviated from the

Networking and Career Success
14
normal distribution. In addition, we tested whether the relative weights differed significantly
from zero. As Johnson (2004) notes, relative weights are proportions and thus confidence
intervals around relative weights will never include zero. To test the significance of a relative
weight Tonidandel, LeBreton, and Johnson (2008) suggest adding a random variable to the
bootstrapping procedure and then test for significant differences between substantive relative
weights and the relative weight of the random variable. For this test, we constructed
confidence intervals of the difference between each substantive and the random relative
weight. If a confidence interval includes zero, the difference is not significant. For this
analysis, we conducted an additional bootstrap with 1000 samples and included a random
variable. As we tested six substantive weights against the weight of the random variable, we
used a Bonferroni correction for our one-sided tests (overall α = .05; α per comparison =
.008).
Results
Table 1 reports means, standard deviations, reliabilities, and correlations among the
variables. Correlations between networking subscales vary between .15 and .60 with a
median correlation of r = .34. Table 1 also shows that four of the six networking scales,
referring to building and maintaining internal as well as external contacts, are significantly
related to salary and career satisfaction at most survey waves. The other two scales, using
either internal or external contacts, are not substantively related to career success indicators.
Table 2 depicts multivariate results for the multilevel regression of salary on
networking to test hypotheses 1 and 2. We examined four hierarchically nested models
following suggestions by Bryk and Raudenbush (2002). Model 1 is the unconditional model
with parameters for the level 1 intercept and slope only. The significant fixed effect for the
slope indicates an increase in salary over measurement occasions. The significant random
effect of the slope indicates individual differences in trajectories of salary growth over time.

Networking and Career Success
15
In model 2, we add control variables. Further, we add the effect of networking in Models 3
and 4. Model 3 tests the effect of networking on concurrent salary. Adding the six scales as a
variable set leads to an improvement in model fit above control variables as indicated by the
significant reduction in deviance, showing support for hypothesis 1. Table 2 also shows that
maintaining external contacts has a significant effect on concurrent salary.
Table 4 shows the relative weights according to Johnson (2000) as well as the
regression weights of the six networking scales, when each scale was added to the control
variables alone. This latter coefficient reflects the contribution of a variable when redundancy
between predictors is ignored and it is shown here for comparison purposes. With regard to
these coefficients, four networking scales, referring to building and maintaining contacts,
have a significant impact on concurrent salary when entered into the multilevel regression
alone. Maintaining external contacts (RW = 45%) and building internal contacts (RW = 24%)
obtained the highest relative weights. Results from the Bootstrap procedure show that all
weights are significantly different from zero. This indicates that all six networking scales
contribute to the significant effect of the scales as a variable set. However, note that the
relative weights for using internal as well as external contacts are also significant, even
though their regression weights are not significant when they are entered into the multilevel
equation alone (cf. Table 4). Also, the bivariate correlations between the two using contacts
scales and salary are not significant. This indicates a potential suppressor relation between
these two and the remaining networking subscales. Also note that confidence intervals for
relative weights overlap considerably indicating that we can not establish significant
differences between substantive relative weights.
To test hypothesis 2, we added the effect of networking on salary growth in model 4
of Table 2. In support of hypothesis 2, model fit improves significantly when the networking
scales are entered as a variable set. Parameter estimates show that maintaining internal

Networking and Career Success
16
contacts has a significant positive impact on salary growth, obtained the highest relative
weight (RW = 49%, see Table 4), and differs significantly from zero.
Hypothesis 3 states that networking is related to concurrent career satisfaction and
hypothesis 4 states that networking is also related to the growth of career satisfaction. To
examine these hypotheses, we calculated three models (see Table 3). Model 1 is again the
unconditional model containing parameters for the intercept and slope only. The slope
parameter (b = -0.031) is not significantly different from zero, showing that career
satisfaction remains stable across time. Moreover, the slope variance is not significantly
different from zero, indicating no individual differences in trajectories of career satisfaction.
In model 2, control variables are entered and we test the effect of networking scales on
concurrent career satisfaction in model 3. The significant difference in deviance between
models 2 and 3 indicates that entering the networking scales as a variable set improve model
fit. Thus, hypothesis 3 is supported. Regression parameters show that maintaining internal
contacts has a significant effect on concurrent career satisfaction. Relative weights as well as
the insignificant regression coefficients of external networking scales when they were entered
alone into a multilevel model (cf. Table 4) indicate that internal networking is of more
importance in predicting career satisfaction. According to the bootstrap procedure, relative
weights for all networking scales, but using external contacts are significantly different from
zero. Hypothesis 4, predicting a growth effect of the networking scales is not supported.
Adding the growth effect did not improve model fit (Δ deviance [6] = 5.3; p > .10). This is
also evident from the insignificant slope variance in the models that indicates the absence of
differential growth trajectories. The growth model is therefore not shown in Table 3. As
career satisfaction remains stable over time, we find no support for hypothesis 4.

Networking and Career Success
17
Discussion
The present study is the first to examine the effects of networking on career success
using a longitudinal research design. We found that networking is related to concurrent salary
level, replicating prior findings (e.g., Forret & Dougherty, 2004). Going beyond prior studies,
our results suggest that networking behaviors can contribute to differential salary growth over
time. In line with the practitioner literature, networking can be considered an investment that
pays off in the future.
Networking was also positively related to concurrent subjective career success, again
replicating prior findings (e.g., Forret & Dougherty, 2004; Langford, 2000). Individuals who
engage in networking behaviors are more satisfied with their careers. Our results further
indicate that internal networking seems to be of higher importance for career satisfaction than
external networking. As career satisfaction remained stable over time, we were unable to find
an effect of networking on changes in career satisfaction. While this finding is not in line
with our assumptions, we believe it is of interest in itself. This unexpected finding shows
some similarities to research on both job and life satisfaction. Studies have shown that
satisfaction is related to stable, dispositional characteristics, such as core self evaluations or
negative and positive affectivity (Diener & Lucas, 1999; Dormann, Fay, Zapf, & Frese,
2006). If career satisfaction is also in part determined by dispositional characteristics,
changes may be more difficult to detect. In a similar vein, set point theory of life satisfaction
suggests that individuals possess a specific level of satisfaction (i.e., the set point) that
remains relatively stable over time. While events may lead to a change of this set point, many
of these changes are temporary and these events lose their impact after three to six months.
Only dramatic events such as unemployment alter the set point (Fujita & Diener, 2005).
Further research should examine whether this theory also applies to career satisfaction and
which events lead to enduring changes of career satisfaction.

Networking and Career Success
18
Theoretical and practical implications
The present study shows that longitudinal designs provide important insights into the
relationship between networking and career success. An intriguing finding is that networking
scales were differentially related to concurrent salary level and salary growth. Our analyses
indicated that all six networking scales were important in the prediction of concurrent salary,
whereas only maintaining internal contacts was an important predictor of salary growth. This
might indicate that individuals with higher salaries can be expected to network as a part of
their job requirements and may in fact point to the possibility that some reverse causation
exists, that individuals may have to resort to specific networking behaviors in order to
accomplish their job. These results also suggest that even though building and using contacts
are essential parts of networking, individuals are well advised to maintain their (internal)
contacts in order to reap the benefits of these acquired contacts in the future. A strong focus
on building contacts may lead to many superficial contacts, but may fail to establish relations
with a minimum amount of trust that is necessary to obtain resources from these contacts. A
focus on using contacts may provide benefits at present, but concurrent use may already be
reflected in concurrent salary and be therefore of less importance for the subsequent
progression of salary growth.
Our findings concerning the importance of maintaining internal contacts for salary
growth may also qualify results obtained by Forret and Dougherty (2004). These authors did
not find a relationship between concurrent salary and their networking scale socializing,
which is comparable to our maintaining internal contacts scale (i.e., theirs includes attending
organizational social functions, going out for drinks after work). While these authors discuss
that socializing may be mainly directed to “peers who tend to have little influence on one’s
compensation” (Forret & Dougherty, 2004 p. 431), our findings indicate that in spite of the

Networking and Career Success
19
lack of an effect of socializing on the concurrent salary level, an effect on salary growth
might nevertheless exist.
Using external contacts had in sum the weakest importance for career success in our
analyses and the significant relative weight indicates that it may even act as a suppressor in
the relation of networking with concurrent salary. A possible explanation for this finding is
that the frequent use of external contacts can be interpreted as a lack of competence, which
might pose a threat to an individual’s reputation. This suggestion might have to be qualified
with regard to particular resources. It may be especially valid for individuals who often seek
task advice and may not apply to strategic information that individuals seek from their
external contacts (Podolny & Baron, 1997). As our networking scales do not distinguish
between the types of resources obtained future research should investigate whether this
assumption is viable.
Future research should also attempt to delineate exactly how networking enhances
career success. The present research has shown that networking leads to salary increases, and
other research (e.g., Thompson, 2005) has shown that networking leads to higher
performance ratings by supervisors. However, it remains unclear whether these outcomes are
achieved by higher work performance or, for example, higher skills in impression
management. Theory on the resources attainable by networking points to both mechanisms
(Wolff et al., in press): As networking yields task related support, it should in turn enhance
work performance and thus performance ratings and salary. However, higher performance
ratings can also be due to higher reputation and higher power as a result of networking. Also,
future research should attempt to assess the joint contribution of individual level networking
behavior and structural level social capital on career success. Social capital and networking
may possess distinct contributions to career success, or social capital may be a mediator of

Networking and Career Success
20
the relationship between networking and career success. Reverse causality is also a plausible
mechanism, e.g., the social capital an individual has acquired may in turn ease networking.
Additionally, future research might consider the opportunities individuals have due to
their life situation outside of work. The present study focused on the work domain, but family
duties such as caring duties for children or elder relatives may also influence networking
behavior, e.g., individuals might have to forgo an opportunity to have drinks after work
because they have to take care of their children. While we controlled for relationship status,
other variables from the family domain might function as confounders or suppressors of the
relation between networking and career success. Likewise, while networking pays off with
regard to career success, costs may be incurred in the family domain, e.g., individuals may
not have much time for their children or may rely on a non-working spouse in order to
network outside their working hours.
As our research underscores the potential benefits of networking, the present findings
may also be useful for career counseling and coaching. For example, conceptualizations of
protean careers suggest that the responsibility to manage a career has shifted from a
predominantly organizational responsibility to the responsibility of individuals (e.g., Hall,
1996). Hall and others have suggested that networking is one means by which individuals can
shape their own careers (Forret & Dougherty, 2004; Sturges et al., 2005) and the present
findings lend further support to this assumption. Employees are well advised to maintain their
internal contacts. It is noteworthy internal networking seems to be of higher importance than
external networking in furthering one’s career.
Limitations
The present study also has some limitations. First of all, even though our longitudinal
design provides further evidence that networking leads to increases in salary, we cannot

Networking and Career Success
21
prove a causal relation between the two variables. Alternatively, third variables may
influence networking as well as career success. By controlling for potentially confounding
variables, e.g., education or job experience, we eliminated the effect of several alternative
explanations. A related concern is that we focus on one measurement of networking at Wave
1 to predict career success, but do not consider networking at subsequent waves. Arguably,
networking behaviors change over time even though empirical findings show that networking
is stable over time (i.e., Sturges, Guest, Conway, & Davey, 2002 report a one year stability of
rtt = .56 that amounts to rtt = .76 corrected for unreliability). Changes in networking behavior,
e.g., by training of networking skills, might have influenced career success. In this vein, our
analyses provide conservative estimates of the effect of networking on career success because
the impact of changes in networking is not taken into account. This argument also highlights
the importance of choosing the right time frame to observe the effects of networking. We
suggest that it takes some time to convert networking behavior into career success, and thus
networking at Wave 1 is of major importance. We thus have provided further, albeit not
definite, evidence for the link between networking and career success.
Second, with regard to the effects of the networking subscales on career success, our
study should be replicated. Shieh and Fouladi (2003) note that parameters of correlated
predictors show sample to sample variation in multivariate analyses that may limit the
generalizability. However, note that these generalizability concerns are limited to the effect of
specific scales, but not to networking scales as a variable set. Also, our additional analyses
using relative weights do shed some further light on the importance of the networking scales
and we consider their use a strength of the present study. A comparison of the significance of
regression weights and relative weights indicates that multicollinearity may indeed result in
too conservative estimates of the importance of correlated predictors. In a similar manner, our
discussion of the importance of maintaining and using internal contacts must be considered in

Networking and Career Success
22
the light of the economic recession in Germany during the time of the study. From 2001 to
2003 unemployment rates rose from 9.4% to 10.5% and the number of job openings
decreased by roughly 30%, from 507,141 to 350,762 (Statistisches Bundesamt, 2004). It is
possible that respondents who focused on maintaining their internal contacts might have had
better chances of increasing their salary due to the decreasing availability of external job
opportunities. Therefore, our study may underestimate the benefits of external networking
behaviors, especially in times of economic upturns. Also, the study has been conducted in
Germany and results may reflect cultural specifics. In Germany, employment security is
comparably high, as federal legislation restricts dismissal of employees to a higher extent
than, for example, in the US. It is possible that Germans are therefore less inclined to build
and maintain external contacts to enhance their career success and focus on internal
networking to a higher extent. In addition, although the networking measures are based upon
theories from international research, our scales might be considered emic (as opposed to etic,
see Brislin, 1976) measures of networking and the particular networking behaviors we assess
may not enhance career success in other cultures. For example, Bozionelos and Wang (2007)
did not find a relationship between a (European) measure of network resources and career
success in China, in spite of the strong emphasis on informal Guanxi relations in the Chinese
culture. Future studies should thus attempt to replicate our results in different cultures and
examine the contingent value of specific networking behaviors.
A third limitation is that we did not include information on career transitions, such as
promotions or employer changes, in our analyses. We therefore do not have information
concerning whether increases in salary were in part achieved by promotions and/ or changes
of employer. While these career transitions describe important steps of the career ladder, their
analysis is not without problems. For example, promotions have often been used as a measure
of career success (see Ng et al., 2005), but they raise problems of comparability (cf. Judge et

Networking and Career Success
23
al., 1995), as respondents in our sample came from a variety of firms and industries. The
meaning of a promotion is influenced by a variety of firm specific factors, e.g., yearly
promotions are common in many consulting firms, or firm size can influence the
opportunities for promotion. Therefore, salary is considered a “better measure of objective
success than the number of promotions because the latter variable is partly confounded with
organizational structure and unmeasured mobility patterns” (Judge et al., 1995 p. 511). In a
similar manner, job changes can occur for a variety of reasons. Open ended comments from
respondents in our sample showed that reasons for a change of employer were mostly related
to career progress (e.g., new challenge, more responsibility), but also to other factors, e.g.,
child birth, relocation of spouse, layoffs, or firm bankruptcy. We chose salary, the most
frequently used indicator of objective career success, because it better reflects the economic
value an organization assigns to employees and their performance. In addition, salary growth
incorporates effects of career transitions that are accompanied by an increase in salary, i.e., if
a promotion or change of employer is accompanied by a pay raise, this effect is also reflected
in our salary measure.
To summarize, this study’s goal was to forge a better understanding of the
relationship between networking and career success. We showed that networking is not only
related to concurrent salary and career satisfaction, but also to salary growth over time. Our
study also suggests that a closer examination of temporal changes in career satisfaction is
advisable.

Networking and Career Success
24
References
Adler, P. S., & Kwon, S.-W. (2002). Social capital: Prospects for a new concept. Academy of
Management Review, 27, 17-40.
Arthur, M. B., Khapova, S. N., & Wilderom, C. P. M. (2005). Career success in a
boundaryless career world. Journal of Organizational Behavior, 26, 177-202.
Becker, T. E. (2005). Potential problems in the statistical control of variables in
organizational research: A qualitative analysis with recommendations. Organizational
Research Methods, 8, 274-289.
Bozionelos, N., & Wang, L. (2007). The relationship of mentoring and network resources in
the Chinese organizational environment. International Journal of human resource
Management, 17, 1531-1546.
Brass, D. J. (1984). Being in the right place: A structural analysis of individual influence in
an organization. Administrative Science Quarterly, 29, 518-539.
Brislin, R. W. (1976). Comparative research methodology: Cross-cultural studies.
International Journal of Psychology, 11, 213-229.
Bryk, A. S., & Raudenbush, S. W. (2002). Hierarchical linear models (2nd ed.). Thousand
Oaks, CA: Sage.
Burt, R. S. (1992). Structural holes. Cambridge, MA: Harvard University Press.
Carroll, G. R., & Teo, A. C. (1996). On the social networks of managers. Academy of
Management Journal, 39, 421-440.
Cohen, A. R., & Bradford, D. L. (1989). Influence without authority: On the social networks
of managers. Organizational Dynamics, 17, 5-17.
Cohen, J., Cohen, P., West, S. G., & Aiken, L. S. (2003). Applied multiple
regression/correlation analysis for the behavioral sciences. (3rd ed.). Mahwah, NJ:
Erlbaum.

Networking and Career Success
25
Coleman, J. S. (1988). Social capital in the creation of human capital. American Sociological
Review, 94 (Suppl.), 95-120.
Curran, P. J. (2003). Have multilevel models been structural equation models all along?
Multivariate Behavioral Research, 38, 529-569.
Diener, E., & Lucas, R. E. (1999). Subjective well-being: Three decades of progress.
Psychological Bulletin, 125, 276.
Dormann, C., Fay, D., Zapf, D., & Frese, M. (2006). A state-trait analysis of job satisfaction:
On the effect of core self-evaluations. Applied Psychology: An International Review,
55, 27-51.
Forret, M. L., & Dougherty, T. W. (2001). Correlates of networking behavior for managerial
and professional employees. Group and Organization Management, 26, 283-311.
Forret, M. L., & Dougherty, T. W. (2004). Networking behaviors and career outcomes:
Differences for men and women? Journal of Organizational Behavior, 25, 419-437.
Frese, M., Fay, D., Hilburger, T., Leng, K., & Tag, A. (1997). The concept of personal
intiative: Operationalization, reliability, and validity in two German samples. Journal
of Occupational and Organizational Psychology, 70, 137-161.
Fujita, F., & Diener, E. (2005). Life satisfaction set point: Stability and change. Journal of
Personality and Social Psychology, 88, 158-164.
Gattiker, U. E., & Larwood, L. (1989). Career success, mobility and extrinsic satisfaction of
corporate managers. Social Science Journal, 26, 75-92.
Goodman, J. S., & Blum, T. C. (1996). Assessing the non-random sampling effects of subject
attrition in longitudinal research. Journal of Management, 22, 627-652.
Greenhaus, J. H., Parasuraman, S., & Wormley, W. M. (1990). Effects of race on
organizational experiences, job performance evaluations, and career outcomes.
Academy of Management Journal, 33, 64-86.

Networking and Career Success
26
Hall, D. T. (1996). Protean careers of the 21st century. Academy of Management Executive,
10, 8-16.
Johnson, J. W. (2000). A heuristic method for estimating relative weight of predictor
variables in multiple regression. Multivariate Behavioral Research, 35, 1-19.
Johnson, J. W. (2004). Factors affecting relative weights: The influence of sampling and
measurement error. Organizational Research methods, 7, 283-299
Judge, T. A., Cable, D. M., Boudreau, J. W., & Bretz, R. D. (1995). An empirical
investigation of the predictors of career success. Personnel Psychology, 48, 485-519.
Kaplan, R. E. (1984). Trade routes: The manager's network of relationships. Organizational
Dynamics, 37-52.
Katz, D., & Kahn, R. L. (1978). The social psychology of organizations (2nd ed.). New York:
Wiley.
Kram, K. E., & Isabella, L. A. (1985). Mentoring alternatives: The role of peer relationships
in career development. Academy of Management Journal, 28, 110-132.
Langford, P. H. (2000). Importance of relationship management for the career success of
Australian managers. Australian Journal of Psychology, 52, 163-168.
LeBreton, J. M., Hargis, M. B., Griepentrog, B., Oswald, F. L., & Ployhart, R. E. (2007). A
multidimensional approach for evaluating variables in organizational research and
practice. Personnel Psychology, 60, 475-498.
Little, T. D., Cunningham, W. A., Shahar, G., & Widaman, K. F. (2002). To parcel or not to
parcel: Exploring the question, weighing the merits. Structural Equation Modeling, 9,
151-173.
Luthans, F., Rosenkrantz, S. A., & Hennessey, H. W. (1985). What do successful managers
really do? An observation study of managerial activities. Journal of Applied
Behavioral Science, 21, 255-270.

Networking and Career Success
27
Maas, C. J. M., & Snijders, T. A. B. (2003). The multilevel approach to repeated measures
for complete and incomplete data. Quality and Quantity, 37, 71-89.
Menard, S. (1991). Longitudinal research. Newbury Park: Sage.
Michael, J., & Yukl, G. (1993). Managerial level and subunit function as determinants of
networking behavior in organizations. Group and Organization Management, 18,
328-351.
Ng, T. W. H., Eby, L. T., Sorensen, K. L., & Feldman, D. C. (2005). Predictors of objective
and subjective career success: A meta-analysis. Personnel Psychology, 58, 367-408.
Nierenberg, A. R. (2002). Nonstop networking. Hendon, VA: Capital Books.
Orpen, C. (1996). Dependency as a moderator of the effects of networking behavior on
managerial career success. Journal of Psychology, 130, 245-248.
Podolny, J. M., & Baron, J. N. (1997). Resources and relationships: Social networks and
mobility in the workplace. American Sociological Review, 62, 673-693.
Raudenbush, S. W. (2001). Comparing personal trajectories and drawing longitudinal
inferences from longitudinal data. Annual Review of Psychology, 52, 501-525.
Schafer, J. L., & Graham, J. W. (2002). Missing data: Our view of the state of the art.
Psychological Methods, 7, 147-177.
Seibert, S. E., Kraimer, M. L., & Liden, R. C. (2001). A social capital theory of career
success. Academy of Management Journal, 44, 219-237.
Shieh, Y.-Y., & Fouladi, R. T. (2003). The effect of multicollinearity on multilevel analysis
parameter estimates and standard errors. Educational and Psychological
Measurement, 63, 951-985.
Snijders, T. A. B., & Boskers, R. J. (1999). Multilevel analysis. London: Sage.
Statistisches Bundesamt. (2004). Statistisches Jahrbuch Deutschland 2004 [Statistical
yearbook Germany 2004]. Wiesbaden, Germany: Statistisches Bundesamt.

Networking and Career Success
28
Sturges, J. J. L., Conway, N., Guest, D., & Liefhooghe, A. (2005). Managing the career deal:
The psychological contract as a framework for understanding career management,
organizational commitment and work behavior. Journal of Organizational Behavior,
26, 821-838.
Sturges, J. J. L., Guest, D., Conway, N., & Davey, K. M. (2002). A longitudinal study of the
relationship between career management and organizational commitment among
graduates in the first ten years at work. Journal of Organizational Behavior, 23, 731-
748.
Thompson, J. A. (2005). Proactive personality and job performance: A social capital
perspective. Journal of Applied Psychology, 90, 1011-1017.
Tonidandel, S., Le Breton, J. M., Johnson, J. W. (2008). Determining the statistical
significance of a relative weight. Manuscript submitted for publication.
Torres, N. L. (2005). It's who you know. Entrepreneur, 33, 128-132.
Wanberg, C. R., Kanfer, R., & Banas, G. T. (2000). Predictors and outcomes of networking
intensity among unemployed job seekers. Journal of Applied Psychology, 85, 491-
503.
Welch, M. S. (1980). Networking. New York: Harcourt Brace Jovanovich.
Willett, J. B. (1997). Measuring change: What individual growth modeling buys you. In E.
Amsel & K. A. Renninger (Eds.), Change and development: issues of theory, method,
and application (pp. 213-243). Mahwah, NJ: Erlbaum.
Witt, P. (2004). Entrepreneurs' networks and the success of start-ups. Entrepreneurship &
Regional Development, 16, 391-412.
Wolff, H.-G., & Moser, K. (2006). Entwicklung und Validierung einer Networkingskala
[Development and validation of a networking scale]. Diagnostica, 52, 161-180.

Networking and Career Success
29
Wolff, H.-G., Moser, K., & Grau, A. (in press). Networking: Theoretical foundations and
construct validity. In J. Deller & D. S. Ones (Eds.), Personality at Work. Mering,
Germany: Hampp.

Networking and Career Success
30
Footnotes
1 The full scale is available from the first author upon request.
2 The reliability estimate of the maintaining internal contacts scale fell slightly below
the “magic threshold” of 0.70. As Cronbach’s alpha for this subscale was above this threshold
in previous studies by Wolff and Moser (2006, i.e., α = .75, α = .71,α = .73 in three studies,
respectively) and possessed adequate stability, we assume that the true reliability of this scale
is close to the threshold of .70 and attribute this minor deviation to sampling fluctuation.
3 This results in two noteworthy consequences. First, the intercept captures the salary
level at Wave 1 and the slope captures changes that occur over later waves. Second, since we
used the natural logarithm of salary as our dependent variable, a linear effect of time on log
salary implies exponential change in salary over time. To investigate the effect of this
implication, we also estimated the models for salary using log time, which implies a linear
effect on salary. Comparing the unconditional models, we found that the linear effect of time
provided the best fit to the data. We therefore decided to use this latter coding.
4 Note that the calculation of relative weights has been described by Johnson (2000)
for ordinary least squares regression, but is also possible for multilevel analysis (Johnson,
personal communication, May 29th, 2008; LeBreton personal communication, May 28th,
2008). Johnson suggests using his approach in structural equation modeling (SEM) and our
multilevel models can be depicted as SEM models (Curran, 2003). The level 2 coefficients of
the networking scales represent fixed effects that are similar to path coefficients in SEM and
OLS regression coefficients (Curran, 2003; Willet, 1997). Even though Johnson (2000) as
well as LeBreton et al. (2007) emphasize the relation of relative weights to R² in OLS
Regression, calculations do neither include nor rely on the unequivocal existence of an R²
measure.

Networking and Career Success
31
Table 1
Means, standard deviations, and correlations of study variables.
M
SD
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
1. GenderA
1.37
0.48
-
2. EducationB
1.43
0.50
-.08
3. Partnership statusC
1.40
0.49
.18**
-.11
4. Job experience
6.74
5.24
-.06
-.19**
-.17**
5. Tenure
5.64
5.44
-.02
-.08
-.15*
.56**
.
6. Supervisor
functionD
0.46 0.50 -.19
*
.10 -.14
*
.15
*
.12
7. Organizational
sizeE
2.02
0.79
-.09
.16*
.06
-.04
-.03
-.06
Networking
8. Internal building
2.45
0.61
-.03
.13
-.05
-.06
-.01
.16*
-.05
(.76)
9. Internal
maintenance
2.65
0.46
-.04
.21**
-.13
-.03
.01
.11
-.11
.32**
(.69)
10. Internal using
2.40
0.49
.08
.08
.09
-.18*
-.14*
-.08
.11
.21**
.35**
(.75)
11. External building 1.95 0.59 -.13 .18
**
-.12 .02 .03 .16
*
.04 .47
**
.34
**
.23
**
(.82)
12. External
maintenance
2.06
0.51
-.06
.19**
-.08
-.07
-.15*
.04
.10
.42**
.42**
.46**
.60*
(.76)
13. External using
1.89
0.47
.05
.05
.09
-.18**
-.24**
-.12
-.06
.15*
.22**
.46**
.23**
.47**
(.76)

Networking and Career Success
32

Networking and Career Success
33
M
SD
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
career satisfaction
14. T1
3.65
0.69
-.10
.09
-.07
.02
.04
.35**
.07
.28**
.21**
.07
.11
.13
.03
(.84)
15. T2
3.63
0.68
-.08
.06
-.11
.16
.09
.40**
.04
.27**
.23**
.03
.16
.12
.01
.70**
(.84)
16. T3
3.62
0.72
.02
.06
-.02
.10
.09
.41**
.07
.22**
.36**
.16
.22**
.24**
.08
.64**
.66**
(.84)
Annual Salary (€)
17. 2000
50498.58
43817.21
-.10
.29**
-.20**
.17**
.12
.26**
.13
.17*
.24**
.02
.32**
.29**
-.05
.26**
.24**
.32**
18. 2001
58907.42
57830.55
-.03
.24**
-.21**
.38**
.26**
.26**
.10
.17*
.32**
.10
.26**
.30**
-.09
.34**
.32**
.39**
.83**
19. 2002
63326.82
65175.79
-.02
.21**
-.20**
.33**
.25**
.23**
.08
.17**
.33**
.12
.23**
.26**
-.04
.30**
.32**
.36**
.74**
.97**
20. 2003
75198.74
94881.97
.02
.14
-.17
.28**
.26**
.22**
.06
.17
.39**
.17
.28**
.34**
-.03
.30**
.30**
.36**
.77**
.91**
.95**
Note: 129 < N < 235. T1 = Time 1; T2 = Time 2; T3 = Time 3.
A 1 = female, 2 = male.
B 1 = no college education, 2 = college education.
C 1 = in steady relationship 2 = single.
D 0 = no 1 = yes.
E 1 = 1 – 500 empl., 2 = 501 – 10000 empl., 3 = more than 10000 employees.
* p < .05.
** p < .01.

Networking and Career Success
34
Table 2
Effects of Networking on Salary.
Model 1
Model 2
Model 3
Model 4
Fixed effects
B
Se
B
Se
B
Se
B
Se
Intercept (β
00
)
10.582**
.034
9.956**
.172
9.598**
.243
9.804**
.255
Slope (β
10
)
0.080**
.007
0.087**
.008
0.087**
.008
-0.062
.054
Gender (β
01
)
-0.110
.061
-0.097
.058
-0.099
.058
Relationship status (β
02
)
-0.169**
.059
-0.137*
.057
-0.138*
.057
Education (β
03
)
0.433**
.063
0.380**
.061
0.381**
.061
Job Experience (β
04
)
0.025**
.008
0.021**
.008
0.021**
.008
Org. Tenure (β
05
)
-0.005
.007
-0.001
.007
-0.001
.007
Org. Size (β
06
)
0.080*
.033
0.080*
.033
0.078*
.033
Supervisor position (β
07
)
0.184**
.060
0.167**
.058
0.166**
.058
Networking: Concurrent
effects
Internal building (β
09
)
0.089
.051
0.066
.055
Internal maintenance (β
0A
)
0.062
.070
-0.006
.076
Internal using (β
0B
)
-0.127
.069
-0.165*
.074
External building (β
0C
)
-0.041
.062
-0.029
.067
External maintenance (β
0D
)
0.248**
.084
0.303**
.090
External using (β
0E
)
-0.065
.071
-0.071
.076

Networking and Career Success
35
Model 1
Model 2
Model 3
Model 4
B
Se
B
Se
B
Se
B
Se
Networking: Growth effects
Internal building (β
11
)
0.017
.015
Internal maintenance (β
12
)
0.051*
.020
Internal using (β
13
)
0.028
.020
External building (β
14
)
-0.008
.016
External maintenance (β
15
)
-0.042
.023
External using (β
15
)
0.005
.020
Random effects
Level 2
Intercept (τ
00
)
0.234**
.024
0.141**
.018
0.124**
.017
0.124**
.017
Slope (τ
11
)
0.004**
.001
0.004**
.001
0.004**
.001
0.003**
.001
Cov (intercept, slope, τ
01
)
-0.003
.004
-0.002
.004
-0.002
.003
-0.002
.003
Level 1 error (σ²)
0.018**
.002
0.019**
.002
0.019**
.002
0.019**
.002
Deviance (npar)
128.5 (6)
38.8 (13)
17.3 (19)
4.0 (25)
Δ Deviance (df)
99.7** (7)
21.5** (6)
13.3* (6)
Note. N = 235. Dependent variable is natural logarithm of salary.
* p < .05.
** p < .01.

Networking and Career Success
36
Table 3
Effects of Networking on Career Satisfaction.
Model 1
Model 2
Model 3
Fixed effects
B
Se
B
Se
B
Se
Intercept (β
00
)
3.681** .058 3.481** .251 2.368** .354
Slope (β
10
)
-0.032
.025
-0.064*
.029
-0.058
.029
log Salary (β
20
)
0.205
**
.045
0.193
**
.046
Gender (β
01
)
0.085 .084 0.072 .080
Relationship status (β
02
)
0.050
.083
0.052
.080
Education (β
03
)
-0.112
.093
-0.164
.090
Job Experience (β
04
)
0.002 .011 0.005 .010
Org. Tenure (β
05
)
-0.011
.010
-0.011
.010
Org. Size (β
06
)
0.039
.046
0.057
.045
Supervisor position (β
07
)
0.432
**
.083
0.384
**
.080
Networking: Concurrent
effects
Internal building (β
09
)
0.117 .070
Internal maintenance (β
0A
)
0.294**
.097
Internal using (β
0B
)
0.113
.094
External building (β
0C
)
0.019 .083
External maintenance (β
0D
)
-0.221 .116
External using (β
0E
)
0.124
.098

Networking and Career Success
37
Model 1
Model 2
Model 3
B
Se
B
Se
B
Se
Random effects
Level 2
Intercept (τ
00
)
0.361**
.087
0.333**
.098
0.313**
.096
Slope (τ
11
)
0.014
.015
0.022
.018
0.023
.018
Cov (intercept, slope, τ
01
)
-0.026
.033
-0.057
.039
-0.061
.039
Level 1 error (σ²)
0.155**
.019
0.159**
.022
0.157**
.022
Deviance (npar)
900.4 (6)
674.7 (14)
653.0
Δ Deviance (df)
225.7** (6)
21.7** (6)
Note. N = 235.
* p < .05.
** p < .01.

Networking and Career Success
38
Table 4
Regression Coefficients and Relative Importance of the Six Networking Subscales.
Salary
Career
Satisfaction
Concurrent
Growth
Concurrent
β (se)
RW
(95% CI)
β (se)
RW
(95% CI)
β (se)
RW
(95% CI)
Internal building
0.139**
(.046)
24%*
(4% - 62%)
0.016
(.013)
10%
(1% - 46%)
0.163*
(.065)
17%*
(2% -45%)
Internal maintenance
0.140*
(.064)
11%*
(2% - 32%)
0.045**
(.017)
49%*
(7% - 76%)
0.323**
(.086)
49%*
(20% - 73%)
Internal using
0.002
(.059)
8%*
(2% - 35%)
0.031+
(.016)
20%
(2% - 43%)
0.212**
(0.79)
14%*
(2% - 31%)
External building
0.109*
(.048)
8%*
(4% - 21%)
-0.005
(.013)
4%
(2% - 20%)
0.063
(.067)
4%*
(2% - 14%)
External maintenance
0.198**
(.055)
45%*
(7% - 60%)
-0.003
(.016)
14%
(3% - 35%)
0.074
(.080)
6%*
(3% - 23%)
External using
0.029
(.061)
4%*
(1% - 16%)
0.009
(.017)
2%
(1% - 27%)
0.164+
(.083)
10%
(1% - 26%)
Note. N = 235. Regression coefficients represent estimates where each of the six scales was added
by itself to control variables in a multilevel regression model. Columns labeled concurrent and
growth show effects of level 2 regression coefficients on the intercept (concurrent effect) and

Networking and Career Success
39
slope (growth effect), respectively. Estimation Confidence intervals and significance test of
relative weights were conducted using empirical intervals from 1000 bootstrap samples. A
Bonferroni correction was used in testing significance of relative weights (overall α = .05; α per
comparison = .008).
+ p < .10.
p < .05.
** p < .01.