The Effects of Job Embeddedness on Organizational Citizenship, Job Performance, Volitional

Absences, and Voluntary Turnover

Author(s): Thomas W. Lee, Terence R. Mitchell, Chris J. Sablynski, James P. Burton and
Brooks C. Holtom
Source: The Academy of Management Journal, Vol. 47, No. 5 (Oct., 2004), pp. 711-722
Published by: Academy of Management
Stable URL: http://www.jstor.org/stable/20159613 .
Accessed: 15/08/2013 09:20

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp

.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.

Academy of Management is collaborating with JSTOR to digitize, preserve and extend access to The Academy
of Management Journal.

http://www.jstor.org

This content downloaded from 131.91.169.193 on Thu, 15 Aug 2013 09:20:18 AM

All use subject to JSTOR Terms and Conditions
 Academy ofManagement journal
2004, Vol. 47, No. 5, 711-722.

THE EFFECTS OF JOB EMBEDDEDNESS ON ORGANIZATIONAL

CITIZENSHIP, JOB PERFORMANCE, VOLITIONAL ABSENCES, AND
VOLUNTARY TURNOVER

THOMAS W. LEE
TERENCE R. MITCHELL
University of Washington, Seattle

CHRIS J. SABLYNSKI
California State University, Sacramento

JAMES P. BURTON
University of Washington, Bothell

BROOKS C. HOLTOM
Georgetown University

This study extends theory and research on job embeddedness, which was disaggre
gated into its two major subdimensions, on-the-job and off-the-job embeddedness. As
revealed that off-the-job embeddedness was signif
hypothesized, regression analyses
icantly predictive of subsequent "voluntary turnover" and volitional absences,
whereas on-the-job embeddedness was not. Also as hypothesized, on-the-job embed
dedness was significantly predictive of organizational citizenship and job perfor
mance, whereas off-the-job embeddedness was not. In addition, embeddedness mod
erated the effects of absences, citizenship, and performance on turnover. Implications
are discussed.

For over 45 years, management scholars have the decisions to participate and to perform than has
theorized about and investigated the been traditionally thought to exist. Recent theory
empirically
causes of employees' jobs, or and research have suggested new and different
voluntarily leaving
turnover" (Maertz & Cam ways to think about turnover, going beyond a strict
"voluntary employee
pion, 1998). In their classic book, Organizations, focus on an employee decision to participate. Add
March and Simon (1958) provided much of the ing considerable richness have been the work of
theoretical for the psychological re Hulin and associates, on a general withdrawal con
underpinning
search on voluntary turnover. They conceptualized struct 1991); of Lee and associates, on
(e.g., Hulin,
employee turnover as a reflection of an employee's multiple paths for leaving, described in their un
decision to participate in the activities of his or her folding model (e.g., Lee & Mitchell, 1994); and of
organization. They also outlined how such a deci Mitchell and associates, on job embeddedness, a
sion to participate differs in substantial ways from construct including both on- and off-the job causes
a decision to perform. As a result of this conceptu of turnover (e.g., Mitchell, Holtom, Lee, Sablynski,
alization, most research on the participation deci & Erez, 2001). Equally these new ideas
important,
sion has treated the performance decision as a have helped scholars better understand the concep
deliberation. A more thorough tual and empirical links between with
largely independent employee
reading of March and Simon and of other, recent drawal and workperformance.
research a closer link between First, we
suggests, however, This study had two specific purposes.
on job embed
sought to extend theory and research
dedness by demonstrating how its major compo
nents (that is, on- and off-the-job embeddedness)
We thank John Cotton (Marquette University) and Greg
differentially predicted the decision to perform
(or
Bigley (University of Washington) for their suggestions
and and
on a draft and Judith Aptaker (University ofWashington) ganizational citizenship job performance)
for their the decision to participate
(volitional absences and
and Greg Anderson (University of Washington)
with the data. voluntary turnover). Second, we sought to show
help
Susan Jackson served as action editor for this manu how these embeddedness components might be
script. processes through which the decisions to perform

711

This content downloaded from 131.91.169.193 on Thu, 15 Aug 2013 09:20:18 AM

All use subject to JSTOR Terms and Conditions
 712 October
Academy of Management Journal

and to participate could be conceptually and em tualization of withdrawal than is found in most
pirically linked. contemporary turnover research. They advocated
for and empirically demonstrated the validity of a
general withdrawal construct (Hanish & Hulin,
CONCEPTUAL BACKGROUND
1991). More specifically, withdrawal was theorized
March and Simon (1958: Chapters 3 & 4) clearly to include multiple work behaviors occurring se
differentiated between the decisions to perform quentially over time, such as poor de
citizenship,
and to participate. They explained the performance creased job performance, increased absences, and
decision in terms of motivational concepts such as finally leaving. The withdrawing person demon
goals, expectancies, and social control (for in strates "a progression of withdrawal from the very
stance, norms, group pressure, and rewards). In mild and easy to the difficult and decisive" (Hulin,
contrast, they explained the participation decision 1998: 11). These ideas suggest that the decisions
in terms of perceived desirability of movement and to perform and participate are related, with the
perceived ease of movement. Over the years, desir decision to perform the decision to
preceding
ability of movement has come to mean work atti participate.
tudes like job satisfaction or organizational com The research Mc
by Lee, Mitchell, Holtom,
mitment, whereas ease of movement has come to Daniel, and Hill (1999) expresses related ideas in
mean perceived job alternatives or actual unem terms of the unfolding model of turnover, accord
ployment rates. More specifically, most turnover ing to which leaving occurs over time and can
theory has the premise that people leave if they are follow various paths. Some turnover happens
unhappy with their jobs and job alternatives are quickly (for instance, a for
preexisting "script"
available. This focus on dissatisfaction, low com leaving drives an employee to quit in response to
mitment, and prevalent job alternatives dominates some event), and some happens more (for
slowly
the study of voluntary turnover. Although gener instance, accumulated job dissatisfaction leads to a
ally valid, the traditional models have had modest search for alternatives). In addition, many people
success in predicting turnover (e.g., Griffeth, Horn, leave because of discernable events,
precipitating
& Gaertner, 2000), with their variables seldom ex and many of these events occur off the job (a spouse
plaining more than 10 percent of variance. relocates, an unsolicited job offer is received).
New ways to think about turnover may be Thus, specific off-the-job events can precipitate
needed. In this research, we attempted to integrate turnover.

March and Simon's ideas about the links between More recently, Mitchell, Holtom, Lee, Sablynski,
the decisions to perform and to participate with and Erez (2001) focused on why stay rather
people
more recent research on employee withdrawal than on how leave. In
by they particular, they drew
Hulin, Lee, and Mitchell. First, March and Simon attention to the reasons people stay through their
(1958) suggested that withdrawal occurs over time job embeddedness construct. the idea of
Reflecting
and includes more types of participation decisions people's being "situated or connected in a social
than just turnover. They stated, "The motivation to web," embeddedness has
several key aspects: (1)
withdraw factor is a general one that holds for both the extent to which people have links to other
absences and voluntary turnover" (March & Simon, people or activities, (2) the extent to which their
1958: 93). In other words, both absences and turn jobs and communities fit other aspects in their "life
over reflect decisions about participation. Second, spaces," and (3) the ease with which links could be
they suggested that many off-the-job factors are im broken?what they would give up if left their
they
portant determinants of why people stay or leave. present settings. Mitchell and his coauthors called
For instance, as March and Simon wrote, "Families these three dimensions links, fit, and re
sacrifice,
often have attitudes about what jobs are appropri spectively, and they are important both on and off
ate for their members" (1958: 72) and "The integra the job. Thus, one can think of a three by two
tion of individuals into the community has fre matrix that shows six dimensions: links, fit, and
quently been urged by organizations because it sacrifice in an organization and in a community.
offers advantages for public relations and reduces Mitchell and his colleagues (2001) provided ini
voluntary mobility" (1958: 72). Thus, March and tial empirical support for job embeddedness. Draw
Simon theorized that severing participation entails ing on data from a sample of retail employees and
more than dissatisfaction-induced It in another of hospital
leaving. sample employees, they first
volves multiple actions, community, and family. reported that job embeddedness was reliably mea
Furthermore, both on- and off-the-job factors are sured as an aggregated score across their six dimen
important antecedents of employee turnover. sions. Second, aggregated job embeddedness corre
Hulin and associates proposed a broader concep lated with intention to leave and predicted

This content downloaded from 131.91.169.193 on Thu, 15 Aug 2013 09:20:18 AM

All use subject to JSTOR Terms and Conditions
 2004 Lee, Mitchell, Burton, and Holtom 713
Sablynski,

subsequent voluntary turnover. Third, job embed That is, leaving a job may have significant effects
dedness significantly predicted turnover after the on an individual's off-the-job life, especially if he
effects of gender, satisfaction, commitment, job or she has to relocate to find new employment.
search, and perceived alternatives had been con More specifically, people who are embedded in
trolled. Thus, job embeddedness was related to one their communities should want to keep their jobs.
of the major decisions about participation, namely, Mitchell and colleagues (2001) reported, for exam
turnover.
ple, that having (1) a working spouse, (2) children
These findings, and the work of March, Simon, in a particular school, or (3) involvement in com
and Hulin, suggest three main ideas: First, job em munity activities was associated with less turnover.
beddedness can be disaggregated into two major To the extent that absences endanger employment
components: embeddedness (that is, or status, they should be lower for people who are
on-the-job
fit, links, and sacrifice) and off-the-job embedded on- and off-the-job.
ganizational
embeddedness (that is, community fit, links, and Extending our reasoning further, off-the-job em
sacrifice). Second, these two components may have beddedness may be more important to the predic
different effects on indicators of performance and tion of turnover and absences than on-the-job em

participation (absences and turnover). Third, em beddedness when satisfaction and commitment
occurs over time, with a deci are on-the-job are controlled.
ployee withdrawal (which constructs)
sion about performing preceding a decision about First, at least some of an
individual's decisions
participating. about absence and
leaving an organization should
be associated with thoughts and considerations
about what would happen if he or she did not have
Hypotheses a job (a hypothetical future or distal state). These
As conceptualized, job embeddedness reflects thoughts (such as job loss owing to being absent too
employees' decisions to participate broadly and di often [Hulin, 1991]) involve potential disruptions
rectly, and itmoves scholarly attention beyond dis to the individual's community involvement, espe
satisfaction-induced leaving. More aptly, job em cially if relocating were required (March & Simon,
beddedness is a retention (or "antiwithdrawal") 1958). In other words, these thoughts do not nec
construct. Hulin
(1998) never directly measured a essarily involve immediate on-the-job consider
general withdrawal construct, instead inferring it ations but do involve off-the-job considerations.
from the occurrence of multiple work behaviors. If Second, Mitchell and his colleagues (2001) re
embeddedness is indeed a broad-based reten ported higher bivariate correlations between on
job
tion (antiwithdrawal) construct and if it captures a the-job embeddedness and
satisfaction, commit
sizable portion of the "decision to participate," ment, and turnover than between off-the-job
both on- and off-the-job embeddedness should pre embeddedness and satisfaction, commitment, and
dict not only employee turnover, but also other turnover. Thus, the effects of on-the-job embedded
withdrawal behaviors, such as decreasing organiza ness on participation may occur in conjunction
tional citizenship behavior, decreasing perfor with work attitudes like satisfaction and commit
mance, and increasing absence. Further, the ex ment, whereas the effects of off-the-job embedded
plained variance in these withdrawal behaviors ness on participation may be less shaped by atti
should exceed that explained by job satisfaction tudes. In other words, on-the-job embeddedness
and organizational commitment. shares more variance with job attitudes than off
In meta-analyses, Griffith and colleagues (2000) the-job embeddedness does; as a result, the higher
showed that job satisfaction and organizational correlation between on-the-job embeddedness and
commitment significantly related to absences and turnover may be mostly due to effects shared with
that absences significantly predicted turnover. Be job attitudes. The lower (but significant) correlation
cause on-the-job embeddedness correlates to satis between off-the-job embeddedness and turnover
faction, commitment, and turnover (Mitchell et al., may reflect different and new information.
2001), it should predict subsequent absences as When considered together, the two arguments
well. However, the effect of on-the-job embedded made above?(1) people think about the future state
ness on absences and turnover may be reduced to of not having a job and its possible effects on com
zero when researchers control for satisfaction and munity involvement and (2) the correlations be
commitment. Further, the ideas of Hulin,
March, tween on- and off-the-job embeddedness and work
Simon, Lee, and Mitchell about nonwork
factors attitudes differ?lead to the following expectation:
suggest that off-the-job embeddedness predicts ab with the attitudes of satisfaction and commitment
sences and turnover, and it may do so even when controlled, the effects of on-the-job embeddedness
satisfaction and commitment are controlled for. on the decision to participate at work should not be

This content downloaded from 131.91.169.193 on Thu, 15 Aug 2013 09:20:18 AM

All use subject to JSTOR Terms and Conditions
 714 Academy Journal October
of Management

significant, but the effects of off-the-job embedded social exchange (Van Dyne & Ang, 1998), norms of
ness on turnover and absences should remain. That reciprocity (Gouldner, 1960), perceived organiza
is, off-the-job embeddedness adds new information tional support (Rhoades & Eisenberger, 2002), and
about why people are absent or leave. work status congruence (Holtom, Lee, & Tidd,
2002) suggest that people come to feel obligated
Hypothesis 1. After job satisfaction and orga
and want to help persons and organizations that
nizational commitment are statistically con
have helped them.
trolled for, off-the-job embeddedness nega Much of the above
reasoning explicitly involves
tively relates to voluntary turnover and
the effect of on-the-job embeddedness on (in-role)
volitional absences, whereas on-the-job em
job performance and (extra-role) organizational cit
beddedness does not predict these withdrawal
izenship. Most importantly, the attributes of a job
behaviors.
and an organization should be significantly more
On the basis of the reviewed theories (Hulin, salient for the immediate motivation (and decision)
1991; Lee & Mitchell, 1994; March & Simon, 1958; to perform than are off-the-job factors. On-the job
Mitchell et al., 2001), we also believe that the de embeddedness should be more proximal to a deci
cision to perform should be related to job embed sion to perform (as manifested by citizenship be
dedness via motivational effects. Because high on haviors and job performance) than the more distal
the-job embeddedness reflects (1) many links, (2) a decision to participate (as reflected by turnover and
good fit, and/or(3) consequential things that an absences, which involve future states and off-the
employee up by quitting,
gives the motivation to job considerations). That is, employees have to per
perform should
be high. That is, employees with form immediately (or right now), whereas they may
high on-the-job embeddedness will (1) be involved be absent next week or quit next month. Although
in and tied to projects and people, (2) feel they fit off-the-job embeddedness should have an effect on
well in their jobs and can apply their skills, and (3) performance, it should be relatively minor. In par
sacrifice valued things if they quit. Correspond ticular, the saliency of the immediate job and orga
ingly, the motivation to perform should be high. nization supersedes, renders less meaningful, or
(Low motivation should occur when on-the-job em makes less potent the more distal effects of off-the
beddedness is low.) job embeddedness on the decision to perform.
The relationship between job embeddedness and
Hypothesis 2. After job satisfaction and orga
the decision to perform can be further specified. In
nizational commitment are controlled for, on
the last decade, the domain of performance has
the-job embeddedness positively relates to or
been divided into in-role and extra-role (e.g., Wil
ganizational citizenship and job performance,
liams & Anderson, 1991). In-role performance is
whereas off-the-job embeddedness is unrelated
similar to
job-description-based specifications of
to these performance indicators.
performance, whereas organizational citizenship
behavior is part of a larger family of extra-role be In their original meta-analysis, Horn and Griffeth
haviors (Van Dyne, Cummings, & McLean Parks, (1995) reported an estimated rho of .33 between
1995). Most often, citizenship is seen as an employ volitional absences and voluntary turnover, and an
ee's actions that help others better perform their estimated rho of -.19 between job performance and
jobs (for instance, training co-workers) and thereby employee turnover. In their update, Griffeth et al.
enhance organizational effectiveness. (2000) reported a rho of .20 between absences and
Conceptually, the more an individual is job em turnover and a rho as -.15 between
performance
bedded (or socially enmeshed) in an organization, and turnover. From these summary findings, an
the more likely he or
she should be to display enduring conclusion is that increased absence sig
citizenship behaviors. In particular, people may be nals more turnover and that good performance sig
(or linked to one another), and nals less turnover. We propose that on-the-job em
interdependent
acts may be consistent with their feelings of beddedness moderates of absences
the effect on
helpful
comfort (or fit) stemmingfrom being part of that turnover and the effect of job performance on turn
social network. The morean employee fits a job, over. As suggested above, higher on-the-job embed
colleagues, and organization, the more natural it dedness reflects more links, better fit, and more
should be to perform citizenship behaviors. In ad consequential losses if an employee quits. As such,
dition, helping others may be perceived as promot people with higher on-the-job embeddedness
ing others' future helpful acts. Foregoing the oppor should to some extent believe and be concerned
tunity to help other interdependent people may that more volitional absences and lower job perfor
well be seen as a sacrificed opportunity to gain an mance may endanger the status of being employed
owed favor. Indeed, the theory and research on and/or attached to their jobs. Conversely, people

This content downloaded from 131.91.169.193 on Thu, 15 Aug 2013 09:20:18 AM

All use subject to JSTOR Terms and Conditions
 2004 Lee, Mitchell Burton, and Holtom 715
Sablynski,

with lower on-the-job embeddedness should hold site. In December 1998, an employee survey was
this belief and concern to a lesser extent. conducted. Individuals' names, employee num
bers, personal characteristics, job satisfaction, orga
Hypothesis 3a. On-the-job embeddedness
nizational commitment, and on- and off-the-job
moderates the positive effect of volitional ab were
embeddedness assessed with voluntary self
sences on voluntary turnover in such a way
reported measures. In January 1999, unit supervi
that these effects are stronger for higher than sors rated their subordinates' organizational citi
for lower on-the-job embeddedness.
zenship behavior and job performance. Absences
3b. embeddedness and turnover for calendar year 1999 were collected
Hypothesis On-the-job
moderates the negative from company records.
effect of job perfor
mance on voluntary turnover in such a way Surveys were distributed to 1,650 employees in
that these effects are stronger for higher than five separate organizational units. Of the 1,650 sur

for lower on-the-job embeddedness. veys, 839 surveys (51%) were returned. Ten sur
veys were not included in the analyses because
Less theoretical and empirical evidence exists on a
they: (1) lacked signed consent form, (2) were
the relationships among organizational citizenship had no or (4)
illegible, (3) identifying information,
and participation (absences and turnover) than ex were identifiable but blank. our
Thus, sample's
ists for in-role performance. For example, a review data come from 829 employees a 50
and represent
by Podsakoff, MacKenzie, Pain, and Bachrach rate. Next, the immediate
percent response super
(2000) did not report empirical evidence on these visors unit of these 829 subordi
(the managers)
relationships. To our knowledge, only Chen, Hui, nates rated their subordinates'
in-role performance
and Sego (1998) have reported that supervisor and extra-role Of the 829 surveys, match
behavior.
rated organizational behaviors (OCBs)
citizenship ing unit manager surveys were returned for 636
predict subordinates' subsequent turnover. To the individuals our sample, 75.3 per
(76.7%). Within
extent that OCBs constitute a form of performance, cent were women; the overall average age was 34.2
however, our prior arguments in Hypotheses 3a =
years 9.9), and the average tenure with
(s.d. the
and 3b should hold for a moderating role of job was 6.6 = The
organization years (s.d. 5.1). major
embeddedness on the effects of OCBs on turnover
ity of respondents had "some college" (48.3%) or a
and absences as well. B.A. or B.S. degree (25.1%). Statistical comparisons
4a. On-the-job and off-the-job em between the sample and overall population (all em
Hypothesis
ployees within
the operations center) yielded no
beddedness moderate the negative effects of
on volun significant differences in age, gender, and tenure.
organizational citizenship behavior
In addition, no significant differences were found
tary employee turnover in such a way that
on turnover, job performance, and organizational
these effects are stronger for than for
higher across our five organizational
citizenship units.
lower on-the-job embeddedness.
Significant differences were found in absences in
Hypothesis 4b. On-the-job and off-the-job em one of our units; the other four units did not differ
beddedness moderate the negative effect of or in absences. Moreover, the ratings of the 20 super
ganizational citizenship on volitional absences visors whorated only a single subordinate were
in such a way that these effects are stronger for to a random of another 20
compared sample super
higher than for lower on-the-job embedded visors who rated multiple subordinates. No signif
ness. icant differences were found between supervisors'
ratings for one versus more subordinates on citizen
ship or job performance. These comparisons sug
METHODS
gest that sampling bias, although not completely
In early 1998, we contacted, visited, and gained discounted, was not a major
problem.
access to data at a regional operations center of a
large international financial institution. The local
Measures
labor market for this operations center was excep
tionally tight, with unemployment below 3 per Voluntary turnover. Each month of the year fol
cent. In September 1998, the two senior authors lowing administration of our survey, the host orga
conducted a focus group with ten randomly se nization a list of individuals who had
provided
lected employees, who discussed
how this study's voluntarily or involuntarily left the organization.
major variables might embed them in their jobs and One hundred thirty-six individuals were classified
community. From this focus group's information, as voluntary leavers (16.4%), and 12 others were
our surveys were tailored to this research terminated. To verify these lists, we
particular involuntarily

This content downloaded from 131.91.169.193 on Thu, 15 Aug 2013 09:20:18 AM

All use subject to JSTOR Terms and Conditions
 716 October
Academy of Management Journal

tried to contact each person who was classified as a did Mitchell and colleagues (2001), we averaged
voluntary leaver. Seventy-two of the 136 voluntary items for on- and off-the-job embeddedness over
leavers were telephoned during the month follow their three subdimensions into composite scores
their and confirmed their = .84 and
ing quitting voluntary (a's .82, respectively).
leaving; the other 64 individuals could not be commitment was assessed with
Organizational
reached. In the analysis, were coded as 0
stayers eight items from Meyer and Allen's (1997) measure
and leavers as 1. of affective commitment. A sample item is "I enjoy
Volitional absences. Thehost organization pro my organization with outside of
discussing people
vided absenteeism records for the year following it." fob satisfaction was measured with three items.
administration of the survey. The organization clas A sample item is "All in all, I am satisfied with my
sified absences as paid (excused) or unpaid (unex Likert scales were used, and factor
job." Five-point
cused). Because we were concerned with volitional analyses indicated unidimensionality. Averaged
absences, our analysis focused on unpaid absences. were used in the analysis
composites (respective
The total number of monthly =
unpaid hours absent a's .85 and .86).
per employee was observed for the 12 months fol An factor analysis with varimax ro
exploratory
lowing the administration of the survey. We were tation was conducted on all items for
self-reported
able to obtain absentee data for 761 employees. job embeddedness, commitment,
organizational
Because some employees left the organization and job satisfaction. Visual of the scree
prior inspection
to the end of the 12-month observation an a three-factor solution. Items for job
period, plot suggested
average monthly absenteeism figure was calculated satisfaction, commitment, and the
organizational
for all persons. Our unpaid absence data also ex fit and sacrifice dimensions of on-the-job embed
hibited a positive skew and a kurtosis and lacked dedness "loaded" on factor 1 (however, one sacri
normality. In order to achieve better fitting and fice item did not load at all). All items for the fit
more normal distributional a square and sacrifice dimensions of off-the-job embedded
properties,
root transformation was ness loaded on factor 2. The links items for on- and
applied.
Organizational citizenship behavior. The im off-the-job embeddedness loaded on factor 3, ex
mediate supervisor of each survey respondent rated cept for two that did not load at all. Given their
the latter's citizenship behavior on eight items that the loading of all items for
conceptual overlap,
were adapted from the Van Dyne and LePine (1998) satisfaction,commitment, and on-the-job embed
organizational citizenship scale. Response options dedness, fit and sacrifice, onto a single factor was to
were 1 ("never") to 5 ("always"), and a sample item be expected and suggested some evidence for con
is "volunteers to do things that are not required." A vergent validity. The separate factors for off-the-job
total of 632 (76.2%) employees were rated. A factor embeddedness, fit and sacrifice, and for links sug
analysis indicated unidimensionality, and an aver gest some evidence of discriminant validity. (The
was used in the analysis = factor pattern matrix is available
aged composite [a .93). upon request to
Job performance. The participants' unit manag the senior author.)
ers assessed their job performance with the six-item Analyses. Logistic regression equations were cal
scale developed by Williams and Anderson (1991). culated for Hypotheses 1, 3a, 3b, and 4a, and ordi
Its response options were 1 ("never") to 5 ("al nary least squares (OLS) regressions were calcu
ways"), and a sample item is "performs all tasks lated for Hypotheses 2 and 4b. We examined the
that are expected of him or her." Job performance main underlying assumptions of all the statistical
ratings were completed for 632 of the employees tests of hypotheses and found no major violations
(76.2%). A factor analysis indicated unidimension (such as outliers, major deviations from normality,
ality, and an averaged composite was used in the or In the variance in
multicolinearity). particular,
=
analysis [a .92). flation factors for the regressions that contained
Job embeddedness. Although most items corre only "main effects" were all below 3. As expected,
sponded directly to Mitchell and associates' mea however, did emerge when in
multicollinearity
sure of job embeddedness, a few minor edits were teraction terms were entered into the regression
required to fit the measure to the current
sample's analyses.
setting. These changesincorporated unique "en
meshing" opportunities available to the employees RESULTS
within the host organization and its local commu
nity. In addition, additional items emerged from Table 1 shows the means, standard deviations,
the focus group and meetings with representative and correlations for all variables in this study. Off
employees, managers, and upper management. The and on-the-job embeddedness significantly related
Appendix shows all our embeddedness items. As to turnover, citizenship, performance, satisfaction,

This content downloaded from 131.91.169.193 on Thu, 15 Aug 2013 09:20:18 AM

All use subject to JSTOR Terms and Conditions
2004 Lee, Mitchell, Sablynski, Burton, and Holtom 717

TABLE 1
b
Means, Standard Deviations, and Correlations*'

Variable Mean s.d.

1. Voluntary turnover 0.16 0.37

2. Performance (in-role) 4.08 0.65 -.12**
3. OCB (extra-role) 3.07 0.88 -.08* .62***
-in-k * *
4. Volitional absences 3.49 7.43 .22***
5. On-the-job embeddedness 2.67 0.49 -.11** .11** .19*** .01
6. Off-the-job embeddedness 2.88 0.54 -.13*** .10** .11** -.16*** .33***
7. Job satisfaction 3.60 0.84 -.10** .02 .07 .03 .73*** .23***
8. Organizational commitment 2.91 0.71 -.09** .02 .06 .07 71 * * * .22*** .69***

a
n = 805
for column 1 (turnover); n > 809 for all other variables [n
> 623 for
performance and OCB).
b
Column 1 reports point-biserial correlations; all other columns report product-moment correlations (two-tailed tests of significance).
The correlations for volitional absences are based on square-root transformations.
* <
p .05
** <
p .01
< .001
***p

and commitment. Whereas off-the-job embedded reports the results of analyses of the hypothesized
ness did, on-the-job embeddedness did not relate direct effects, shows the regression of turnover onto
to volitional absences. satisfaction, commitment, and on- and off-the-job
significantly
embeddedness. As hypothesized, the coefficient for
off-the-job embeddedness is significant and shows
Tests of Hypotheses
a negative effect, whereas the coefficient for on-the
Hypothesis 1 holds that, when satisfaction and job embeddedness is nonsignificant. Equation 2
commitment are statistically controlled, off-the-job shows the regression of absences onto satisfaction,
embeddedness remains negatively related with commitment, and off- and on-the-job embedded
turnover and absences, whereas on-the-job embed ness. The coefficient for off-the-job embeddedness
dedness is unrelated. Equation 1 in Table 2, which is significant and negative, whereas the coefficient

TABLE 2
Effects of Job Embeddedness

Dependent Variables

Equation 1: Equation 2: Equation 4:

Voluntary Voluntary Equation 3: Organizational

Predictors Turnover3 Absences1* Job Performance1* Citizenship Behavior'1

Job satisfaction .91 .01 -.19** .10

Organizational commitment .92 .11* .06 .11

On-the-job embeddedness .80 -.02 .27*** .32***

Off-the-job embeddedness .60** -.18*** .07 .05

-
F or 2 log-likelihood 706.78 6.78*** 6.24*** 8.52***
R2 .04 .04 .05
AR2 or A*2

On-the-job embeddedness .02* .04*

Off-the-job embeddedness 7.75** .03*

17 805 739 620 620

a
Logistic regression. The entries are b's. Entries above 1.00 indicate effects, and entries below 1.00 indicate
exponentiated positive
negative effects.
b
The entries are standardized regression coefficients when all variables are entered into the equation.
* <
p .05
** <
p .01
< .001
***p
Two-tailed tests.

This content downloaded from 131.91.169.193 on Thu, 15 Aug 2013 09:20:18 AM

All use subject to JSTOR Terms and Conditions
 718 Journal October
Academy of Management

for on-the-job embeddedness is nonsignificant. TABLE 3

Thus, Hypothesis 1 is supported. Moderating Effects of Job Embeddedness and
Hypothesis 2 holds that, when satisfaction and Work Behaviors on Voluntary Turnover3
commitment are controlled, on-the-job embedded
Predictors 1 Equation 2 Equation 3
ness remains positively related with citizenship
Equation

and performance, whereas off-the-job embedded 0.89 0.76

Job satisfaction 0.76
ness is unrelated. Equation 3 in Table 2 shows the 0.83 0.85 0.89
Organizational
regression of performance onto satisfaction, com commitment

and off- and on-the-job embeddedness. embeddedness 0.49 14.27 7.26*

mitment, On-the-job
**
Off-the-job embeddedness 0.68 0.09 0.09
As hypothesized, only the coefficient for on-the-job
Voluntary absences 0.60
embeddedness is significant and positive, whereas 0.89
Job performance
the coefficient for off-the-job embeddedness is not 0.78
Organizational
significant. Equation 4 shows the regression of cit citizenship behavior

izenship onto satisfaction, commitment, and off

and on-the-job embeddedness. The coefficient for
On-the-job embeddedness 1.37*
on-the-job embeddedness is significant and posi X absences
voluntary
tive, whereas the coefficient for off-the-job embed
Off-the-job embeddedness 1.01
dedness is not significant. Thus, Hypothesis 2 is X
voluntary absences

On-the-job embeddedness 0.50*

supported. X
3a predicts that on-the-job embed job performance
Hypothesis embeddedness 1.65
Off-the-job
dedness moderates the positive effect of volitional X
job performance
absences on quitting, and Hypothesis 3b predicts embeddedness 0.49*
On-the-job
that on-the-job embeddedness moderates the nega X
organizational
tive effect of job performance on quitting. In each citizenship behavior

Off-the-job embeddedness 1.92*

case the moderation is such that the effects are
X
organizational
stronger for higher than for lower on-the-job em
citizenship behavior
beddedness. Equation 1 in Table 3, which reports
the results of analyses of the hypothesized moder
ation effects, shows the regression of turnover onto -2 log-likelihood 622.92 505.40 504.80

on- and off-the-job em A#2b 6.05* 4.82+ 11.15***

satisfaction, commitment, n
740 621 621
beddedness, absences, and the interactions be
tween on-the-job embeddedness and absences and a
Logistic regressions. The entries are exponentiated b's. En
between off-the-job embeddedness and absences. tries above 1.00 indicate positive effects, and entries below 1.00

As hypothesized, the coefficient for the interaction indicate negative effects.

b
For interactions.
between on-the-job embeddedness and absences is +
p < .10
statistically significant and shows a positive effect, * < .05
p
whereas the interaction between off-the-job embed **p
< .01
*** <
dedness and absences is nonsignificant. To de p .001

scribe this interaction, we calculated re Two-tailed tests.

separate
gressions for high and low groups based on a
median split of on-the-job embeddedness. The high low groups based on a median split of on-the-job
group has a steeper positive slope than the low embeddedness. The high group has a negative and
on turnover = = <
group for absences regressed (exp b significant slope (exp b 0.41, p .001), whereas
1.63, p < .001, vs. 1.27, p < .01). Equation 2 in the low group has a negative but nonsignificant
Table 3 shows the regression of turnover onto sat slope for performance regressed on turnover. Thus,
isfaction, commitment, on- and off-the-job embed Hypotheses 3a and 3b are supported.
dedness, performance, and the interactions be Hypothesis 4a predicts that on- and off-the-job
tween on-the-job embeddedness and performance embeddedness moderate the negative effect of or
and between embeddedness and perfor on quitting (turnover), and
off-the-job ganizational citizenship
mance. The coefficient for the interaction between Hypothesis 4b makes a similar prediction for voli
on-the-job embeddedness and performance is sig tional absences. In both cases, the moderation is
nificant and shows a negative effect, whereas the such that these effects are stronger for higher than
coefficient interaction
for the between off-the-job for lower on-the-job embeddedness. In Table 3,
embeddedness and performance is nonsignificant. equation 3 shows the regression of turnover onto
To describe the significant interaction, we calcu satisfaction, commitment, on- and off-the-job em
lated separate regression equations for high and beddedness, citizenship, and the interactions be

This content downloaded from 131.91.169.193 on Thu, 15 Aug 2013 09:20:18 AM

All use subject to JSTOR Terms and Conditions
 2004 Lee, Mitchell, Burton, and Holtom 719
Sablynski,

tween embeddedness and citizenship dedness, and off-the-job embeddedness) with the
on-the-job
and between off-the-job embeddedness and citizen variance explained in the corresponding regres
As the coefficients for the in sions containing satisfaction, commitment, and
ship. hypothesized,
teractions between on-the-job embeddedness and overall embeddedness. For the corresponding lo
and between embeddedness gistic regression analyses, we compared d?viances
citizenship off-the-job
and citizenship are statistically To de from these two equations with a G-test (Hosmer &
significant.
scribe these interactions, we calculated separate Lemeshow, 2000). To reiterate, the new regressions
for high and low groups based on a (both the OLS and logistic versions), with their
regressions
median split of both on- and off-the-job embedded coefficients for overall embeddedness, allowed for
ness. For the high on-the-job embeddedness group, stronger inferences on the strength of predictions.
the slope for citizenship on turnover is For absences, the four-variable model explained 4
regressed
b = < of variance = whereas the three
negative and significant (exp 0.55, p .01), percent [R2 .04),
3 percent =
whereas the slope for the low group is nonsignifi variable model explained [R2 .03; both
cant. For the high off-the-job embeddedness group, p < .001). The difference between the two R2s was
on turnover = < a
the slope for citizenship regressed is significant [F 3.20, p .05, with one-tailed test,
nonsignificant, whereas the slope for the low group but p < .10, with a two-tailed test). For perfor
=
is negative and significant (exp b 0.65, p < .05). mance, the four- and three-variable models had the
Thus, Hypothesis 4a is supported. same respective explained variances as the models
For Hypothesis 4b, absences were regressed onto for absences (both p < .001). The difference in ?2s
on- and off-the-job em was - < .05 with one- and
satisfaction, commitment, significant [F 6.01, p
beddedness, and the interactions between on-the two-tailed tests). For citizenship, the four-variable

job embeddedness and citizenship and between model explained 5 percent of variance, and the
embeddedness and (This re three-variable model, 4 (both p < .001).
off-the-job citizenship. percent
not is The difference in fl2s was =
gression is shown but available upon request significant (F 11.51,
to the senior author.) Only the interaction for on p < .001, with one- and two-tailed tests). For turn
the-job embeddedness and citizenship is statisti over, the four-variable model had -2 log-likelihood
=
cally significant (? -0.56, p < .05). To describe of 706.78, whereas the three-variable model had a
this interaction, we calculated separate regression -2 log-likelihood of 707.31. The difference be

equations for high and low groups based on a me tween them (G) was nonsignificant. In sum, these
dian split of on-the-job embeddedness. The high data generally support the stronger inferences.
group showed a negative and significant slope that
was steeper than the low group's = =
(? -0.19, p DISCUSSION
.001, vs. -0.12, p < .05). Thus, Hypothesis 4b is
only partially supported. This study expands understanding of job embed
dedness. First, off-the-job embeddedness predicted
turnover and absences, whereas on-the-job embed
dedness did not (Hypothesis 1). In contrast, on-the
Post Hoc Analyses
job embeddedness predicted organizational citi
Table 2 shows different
patterns of significant zenship and job performance, whereas off-the-job
contribution towardprediction of turnover, ab embeddedness did not (Hypothesis 2). Second, the
sences, performance, and citizenship for on- and two components of job embeddedness may be pro
off-the-job embeddedness. Post hoc, we tested the cesses through which the decisions to perform and
stronger inferences that off-the-job embeddedness to participate can be conceptually and empirically
is a better predictor than on-the-job embeddedness linked. On-the-job embeddedness moderated the
for turnover and absences and that on-the-job em positive effect of volitional absences on turnover
beddedness is a better predictor than off-the-job (Hypothesis 3a), the negative effect of job perfor
embeddedness for performance and citizenship. mance on turnover (Hypothesis 3b), and the nega
First, by adding scores for on- and off-the-job em tive effect of citizenship on absences; the modera
beddedness we created an overall
job embedded tion was such that these effects were stronger for
ness score. Second, we
regressed each dependent higher than for lower on-the-job embeddedness
variable onto the predictors of satisfaction, commit (Hypothesis 4a).
ment, and overall embeddedness. Third, we com Three particular limitations of this study should
pared the difference in variance explained [R2) in be noted. First, the timing for some of our measures
the three OLS regressions shown in Table 2 (that is, provides only limited support for causal infer
predicting absences, performance, and citizenship ences. Although our four behavioral outcome vari
with satisfaction, commitment, on-the-job embed ables were measured independently from and after

This content downloaded from 131.91.169.193 on Thu, 15 Aug 2013 09:20:18 AM

All use subject to JSTOR Terms and Conditions
 720 Journal October
Academy of Management

our respondents' of satisfaction, com that measure

self-reports job embeddedness immediately be
mitment, and embeddedness, organizational citi fore and after acts of organizational citizenship and
zenship and job performance were only a
assessed formal appraisals of in-role performance, absence
month after the self-reported survey. Although it is spells, or individual quitting might yield valuable
likely that job embeddedness was a cause of our evidence on causal effects.
outcome variables, the reverse direction may hold. In our view, the managerial of this
implications
Second, the measures of on- and off-the-job embed study are clear. Job embeddedness (which can be
dedness are still preliminary and evolving. Al established through building community, develop
though our data on factor structures and internal ing a sense of belonging, ties
establishing deep
consistencies produce empirical findings similar to among employees, and deepening social capital)
earlier work, these measures are not yet established may increase retention, attendance, citizenship,
and standard researchinstruments. Third, we dis and job performance. Furthermore, organizations
aggregated embeddedness into on- and off-the job can be proactive about job embeddedness: links can
components. In the future, profiles of the six em be increased through teams and long-term projects;
beddedness dimensions may be useful for predic sacrifice can be increased by connecting job and
tion and understanding. organizational rewards to longevity; and fit can be
It should be mentioned that Meyer and Allen's increased by matching employees' knowledge,
(1997) dimension of continuance commitment and skills, and attitudes with
abilities, a job's require
the embeddedness subdimension of organization ments. Equally important, managers can increase
related sacrifice (sacrifice?organization) are simi off-the-job embeddedness by providing people
lar. Both dimensions share notions of "sunk cost" with information about the community surround
and reluctance to give things up by leaving. How ing their workplace and by providing social sup
ever, the original continuance commitment con port for local activities and events (Mitchell, Hol
struct combined reluctance with the availability of tom, & Lee, 2001).
alternatives, which organization-related sacrifice In closing, our results suggest that studying em
does not. Even if items for alternatives are omitted reasons for both and
ployees' staying leaving may
from the continuous commitment measure, items enrich knowledge of retention, increasing it beyond
for organization-related sacrifice have much more what the current focus on leaving permits. This
specific and targeted referents such as perks, re broader perspective suggests an interesting and po
spect, compensation, benefits
(retirement and tentially fruitful direction for future research.
health care), and promotional opportunities. Thus,
our measure omits the "alternatives" idea and of
fers more detail in terms of the specific REFERENCES
topics than
continuance commitment. (See Yao, Lee, Mitchell, X. &
Chen, P., Hui, C, Seg?, D. J. 1998. The role of
Burton, and Sablynski [2004] for a comprehensive organizational citizenship behavior in turnover:

comparison between job embeddedness and related Conceptualization and preliminary test of key hy
constructs, continuance commitment.) potheses. Journal of Applied Psychology, 83: 922
including
our results indicate the appropriateness 931.
Overall,
of studying retention and performance as tandem Gouldner, A. W. 1960. The norm of A pre
reciprocity:
job behaviors and viewing job embeddedness as a liminary statement. American Sociological Re
meaningful mechanism through which to under view, 25: 161-178.

stand this linkage. In particular, Mitchell and co Griffeth, R. W., Horn, P. W., & Gaertner, S. 2000. A meta
authors (2001) reported significant predictive asso of antecedents and correlates of
analysis employee
ciations between aggregated job embeddedness and turnover:
Update, moderator tests, and research im

turnover in two In this study, disaggre plications for the millennium. Journal of Manage
samples.
embeddedness ab ment, 26: 463-488.
gated job predicted turnover,
sences, in-role performance, and organizational cit Hanish, K. A., & Hulin, C. L. 1991. General attitudes and

izenship. Thus, meaningful statistical effects were organizational withdrawal: an evaluation of a causal

found over three diverse and sizable In model. Journal of Vocational Behavior, 39: 110
samples.
the future, itmay be timely for researchers to move 128.

beyond simple prediction and predictive validity Holtom, B. C, Lee, T. W., & Tidd, S. T. 2002. The rela

designs. Given the existing studies, research de tionship between work status congruence and work

that allow for stronger causal inferences are related attitudes and behaviors. Journal of Applied
signs
now needed. Field or quasi field experiments that Psychology, 87: 903-915.
include interventions aimed at altering job embed Horn, P. W., & Griffeth, R. W. 1995. Employee turnover.
dedness are suggested. Alternatively, field studies Cincinnati: South-Western.

This content downloaded from 131.91.169.193 on Thu, 15 Aug 2013 09:20:18 AM

All use subject to JSTOR Terms and Conditions
 2004 Lee, Mitchell, Burton, and Holtom 721
Sablynski,

Hosmer, D. W., & Lemeshow, S. 2000. Applied logistic Van Dyne, L., Cummings, L. L., &McLean Parks, J. 1995.
regression (2nd ed.). New York: Wiley. Extra-role behaviors: In pursuit of construct and def
C. L. 1998. constructs time: Pot
initional clarity (a bridge over muddied waters). In
Hulin, Behaviors, and
L. L. Cummings & B. M. Staw (Eds.), Research in
holes on the road well traveled. Invited address,
behavior, vol. 17: 215-285. Green
annual meeting of the Society for Industrial and organizational
wich, CT: JAI Press.
Organizational Psychology, Dallas.
Williams, L. J., & Anderson, S. E. 1991. Job satisfaction
Hulin, C. L. 1991. and commit
Adaptation, persistence
and commitment as of cit
ment in In M. D. Dunnette & L. M. organizational predictors
organizations.
izenship and in-role behaviors. Journal of Manage
Hough (Eds.), Handbook
of industrial and organi
ment, 17: 601-617.
zational psychology (2nd ed.): 445-507. Palo Alto,
CA: Consulting Psychologists Press. Yao, X., Lee, T. W., Mitchell, T. R., Burton, J. P., &

Hulin, C. L. 2002. Lessons from industrial and organiza Sablynski, C. S. 2004. Job embeddedness: Current
research and future directions. In R. Griffeth & P.
tional psychology. In J. Brett & F. Drasgow (Eds.),
Horn (Eds.), Understanding employee retention
The psychology of work: Theoretically based em
and turnover: 153-187. Greenwich, CT: Information
pirical evidence. Mahwah, NJ: Erlbaum.
Age.
Lee, T. W., & Mitchell, T. R. 1994. An alternative ap

proach: The unfolding model of voluntary employee APPENDIX

turnover. Academy Review, 19:
of Management
51-89.
Job Embeddedness Itemsa

Lee, T. W., Mitchell, T. R., Holtom, B. C, McDaniel, L., & Fit, community
Hill, J.W. 1999. Theoretical development and exten I really love the place where I live.b
sion of the unfolding model of voluntary turnover. I like the family-oriented environment of my
Academy of Management Journal, 42: 450-462. community.
This community I live in is a good match for me.
Maertz, C. P., & Campion, M. A. 1998. 25 years of volun
I think of the community where I live as home.b
tary turnover research: A review and Inter
critique. The area where I live offers the leisure activities that
national Review of Industrial and Organizational
I like sports, outdoors, cultural, arts).
(e.g.,
Psychology, 13: 49-81.

March, J. G, & Simon, H. A. 1958. Organizations. New Fit, organization

York: Wiley. My job utilizes my skills and talents well.b
I feel like I am a good match for this organization.13
Meyer, J. P., & Allen, N. J. 1997. Commitment in the
I feel personally valued by (name of the
workplace: Theory, research, and application.
Thousand Oaks, CA: Sage. organization).
I like my work schedule (e.g., flextime, shift).
Mitchell, T. R., Holtom, B. C, & Lee, T. W. 2001. How to I fit with this organization's culture.b
keep your best employees: The development of an I like the authority and responsibility I have at this
effective attachment policy. Academy of Manage company.13
ment Executive, 15(4): 96-108.
Mitchell, T. R., Holtom, B. C, Lee, T. W., C. J., Links, community
Sablynski,
& Erez, M. 2001. Why Are married?b
people stay: Using job embed you currently
to If you are married, does your spouse work outside
dedness turnover.
predict voluntary Academy of
44: 1102-1121. the home?b
Management Journal,
Do you own the home you live in? (mortgaged or
Podsakoff, P. M., MacKenzie, S. B., Pain, J. B., & Bach
outright)
rach, D. G. 2000. behav
Organizational citizenship
My family roots are in the community where I live.
iors: A critical review of the theoretical and empiri
cal literature and suggestions for future research.
Links, organization
Journal of Management, 26: 513-564. How long have you been in your present position?
Rhoades, L., & Eisenberger, R. 2002. Perceived (years)b
organiza
tional support: A review of the literature. Journal How long have you worked for this organization?
of
Applied Psychology, 87: 698-714. (years).b
How long have you worked in the (banking)
Van Dyne, L., & Ang, S. 1998. Organizational citizenship
behavior of contingent workers in Singapore. Acad industry? (years).b
How many coworkers do you interact with
emy of Management Journal, 41: 692-703.
regularly?b
Van Dyne, L., & LePine, J. A. 1998. Helping and voice How many coworkers are highly dependent on you?b
extra-role behaviors: Evidence of construct and pre How work teams are on?b
many you
dictive validity. Academy of Management Journal, How many work committees are you on?b
41: 108-119. Continued

This content downloaded from 131.91.169.193 on Thu, 15 Aug 2013 09:20:18 AM

All use subject to JSTOR Terms and Conditions
 722 Academy of Management Journal October

APPENDIX Continued A

Sacrifice, community
Leaving this community would be very hard.b Thomas W. Lee (orcas@u.washington.edu) is a professor
me a lot in my of human resource and be
People respect community.13 management organizational
My neighborhood is safe.b havior and the Evert McCabe Faculty Fellow at the Uni
versity of Washington Business School. He earned his
Sacrifice, organization Ph.D. in organizational studies at the University of Ore
I have a lot of freedom on this job to decide how to gon. His current research interests include employee
re

pursue my goals.b tention, staffing, work motivation, and research methods.

The perks on this job are outstanding.13
Terence R. Mitchell (trm@u.washington.edu) is the Ed
I feel that people at work respect me a great deal.b
ward E. Carlson Professor of Business Administration
Iwould incur very few costs if I left this
and a professor of psychology at the University of Wash
organization.0
ington Business School. He earned his Ph.D. in social
Iwould sacrifice a lot if I left this job.b
psychology at the University of Illinois. His current re
are excellent here.b
My promotional opportunities
search interests include employee retention, work moti
I am well for my level of
compensated performance.13
vation, and decision making.
The benefits are good on this job.b
I believe the prospects for continuing employment Chris J. Sablynski is an assistant
(sablynsk@csus.edu)
with this are excellent.13 of behavior and environment at
company professor organizational
California State University, Sacramento. He earned his
a
Items 1-3 for links, community, and links, organization, Ph.D. in behavior and human resource
organizational
were standardized before being analyzed or included in any
management at the University of Washington Business
composites. School. His current research interests include employee
b
Item used by Mitchell, Holtom, Lee, Sablynski, and Erez retention and deviance.
workplace
(2001).
c is an assistant
Reverse coded. James P. Burton (jburton@uwb.edu) pro
fessor of management at the University of Washington,
Bothell. He earned his Ph.D. in organizational behavior
and human resource management at the of
University
Business School. His current research inter
Washington
ests include fairness, retention, and
workplace employee
teaching effectiveness in university settings.

Brooks C. Holtom (bch6@msb.edu) is an assistant profes

sor of management in the McDonough School of Business
at Georgetown University. He earned his Ph.D. in orga
nizational behavior and human resource at
management
the University of Washington Business School. His cur
rent research interests include human and social capital
development and employee retention.

This content downloaded from 131.91.169.193 on Thu, 15 Aug 2013 09:20:18 AM

All use subject to JSTOR Terms and Conditions