Academic Journal Review

O R I G I N A L P A P E R

Friends and Family: The Role of Relationships in Community and Workplace Attachment

Jorge A. Gonzalez1 • Belle Rose Ragins2 • Kyle Ehrhardt3 • Romila Singh2

Published online: 7 October 2016

� Springer Science+Business Media New York 2016

Abstract

Purpose Most work–life research focuses on the spillover

of the nuclear family to the workplace, offering little

insight into how other family relationships and friendships

can spill over to affect employees’ organizational attach-

ment. Past research has also overlooked the role of rela-

tionship quality and the mechanisms underlying these life-

to-work spillover effects. Addressing these shortcomings,

we integrate the systemic model of community attachment

with job embeddedness theory to develop a model of

community relational embeddedness and then use this

model to examine how nonwork relationships connect

people to their workplaces.

Design/Methodology/Approach We used survey data from

a national sample of 2025 accounting professionals and

tested mediation hypotheses using structural equation

modeling.

Findings Employees’ relationships with friends and family

predicted their attachment to their communities, which in

turn predicted their workplace turnover intentions. Sup-

porting our theoretical model, bonds with friends and

family predicted moving intentions, and community fit and

sacrifice mediated these effects. Community fit and

sacrifice also predicted work turnover intentions indirectly

through moving intentions. Tests also revealed that, sur-

prisingly, friendships had a stronger impact on community

attachment than family.

Implications Employees are connected to their organiza-

tions through an array of close community relationships

that extend beyond the nuclear family (i.e., spouse, chil-

dren). Organizations can enhance employees’ workplace

attachment by recognizing the role of friends and offering

work–life programs that use a broad conceptualization of

family (e.g., adult siblings, parents).

Originality/Value Our study illustrates the importance of

community relationships to workplace attachment, and the

need to incorporate relational quality, nonnuclear family,

and friendships in future research.

Keywords Relationships � Work–life � Community attachment � Job embeddedness � Turnover intentions � Spillover � Family � Friends

Introduction

Although our relationships extend beyond the workplace,

we know little about the role community relationships play

in employees’ attachment to their organizations. Sociolo-

gists have long recognized that a wide range of relation-

ships can attach people to their communities (Kasarda and

Janowitz 1974; Sampson 1988), but organizational scholars

know little about how these relationships attach people to

their workplace. Most work–life research narrowly focuses

on the impact of the nuclear family (i.e., an employee’s

spouse/partner and young children) on the workplace (cf.,

Casper et al. 2007; Eby et al. 2005). However, this fails to

capture the wider constellation of other familial

& Jorge A. Gonzalez [email protected]

1 Department of Management, Robert C. Vackar College of

Business & Entrepreneurship, University of Texas – Rio

Grande Valley, 1201 W. University Dr., Edinburg,

TX 78539, USA

2 Department of Management, Lubar School of Business,

University of Wisconsin – Milwaukee, Milwaukee,

WI 53211, USA

3 Department of Management, Business School, University of

Colorado Denver, Denver, CO 80202, USA

123

J Bus Psychol (2018) 33:89–104

https://doi.org/10.1007/s10869-016-9476-3

relationships (e.g., parents, grandparents, adult siblings,

cousins, in-laws, adult children, and grandchildren) that

can anchor people to their communities and workplaces.

Friendships have also been overlooked in work–family

research, even though they can be a unique and powerful

source of support and attachment (Procidano and Heller

1983; Wellman and Wortley 1990; Wiseman 1986). The

impact of community relationships on the workplace is

therefore an important but relatively neglected area of

inquiry that can inform research on the mechanisms

underlying employee attachment and managers seeking to

retain talented workers.

Accordingly, the purpose of our study is to assess the

role of community relationships in organizational attach-

ment. Organizational scholars increasingly recognize the

important role community experiences play in workplace

attachment and have found that employees are less likely to

leave their work organizations when they are embedded in

their communities (Feldman et al. 2012; Mitchell et al.

2001). Sociologists and community psychologists offer

complementary theoretical insights about the role of

community relationships in community attachment

(Kasarda and Janowitz 1974; Raymond et al. 2010;

Sampson 1988). Marrying these literatures, we offer a

theoretical foundation for examining how relationships

with friends and nonnuclear family attach employees to

their communities and ultimately their workplaces.

Drawing on the sociological literature, we apply the

systemic model of community attachment (Kasarda and

Janowitz 1974; Sampson 1988) to offer the concept of

community relational embeddedness, which we define as

the degree to which nonwork relationships bind people to

their communities. As described later, we use key theo-

retical tenets from the community psychology literature to

develop the construct of relational bonds, which assesses

the quality of employees’ nonwork relationships and their

importance for community attachment (Hidalgo and Her-

nández 2001; Raymond et al. 2010). We then integrate and

apply key constructs from job embeddedness theory

(Mitchell et al. 2001) to examine the intervening mecha-

nisms that link community relationships to the workplace.

This theoretical integration offers a potent frame for our

study by explaining how relationships with family and

friends create community and workplace attachment.

We seek to make several contributions to theory,

research, and practice. First, from a theoretical perspective,

our application of sociological and community psychology

literatures enriches job embeddedness theory and offers a

more substantive approach for understanding the role of

community relationships in organizational life. These the-

oretical perspectives also complement one another: Soci-

ological perspectives describe how relationships with

friends and family in the community affect community

attachment, while job embeddedness theory explains the

mechanisms through which community relationships attach

people to the workplace. Our study also offers a more fine-

grained analysis by empirically distinguishing relationships

with friends and family. This allows us to parcel out and

theoretically understand the distinct effects of friends and

family bonds on employees’ community and workplace

attachment. Taken together, therefore, our research seeks to

offer job embeddedness and work–life scholars new theo-

retically driven insights into the role of community rela-

tionships in organizational life.

Second, our study offers a more complete assessment of

the impact of community relationships on workplace

attachment by incorporating both the presence and quality

of relationships in our measures. As described later, soci-

ologists (e.g., Flaherty and Brown 2010; Kasarda and

Janowitz 1974; Sampson 1988) and community psycholo-

gists (e.g., Crowe 2010) observe that it is important to

assess the proportion of friends and family who live in

one’s community, as people may be more attached to

communities where most of their friends and family reside.

Relationship scholars point out, however, that it is not just

the presence of relationships but also their quality that

matters and that researchers need to incorporate indicators

of relational quality in their analyses (Bengtson 2001;

Carmeli and Gittell 2009; Stephens et al. 2013). Despite

these recommendations, job embeddedness and work–life

scholars have generally examined the presence of rela-

tionships, but not their quality (cf., Eby et al. 2005; Feld-

man et al. 2012; Laschober et al. 2012; Mitchell et al.

2001). This is an important omission, as even if employees

have family members and friends in their communities,

they may not see them often or value these relationships.

We empirically address this issue in our study and, in so

doing, offer a more comprehensive approach for assessing

how community relationships influence organizational

attachment.

Our third contribution speaks to practice. Many orga-

nizations now recognize the importance of ‘‘family

friendly’’ policies and the positive impact such policies can

have on employees’ retention, commitment, job satisfac-

tion, and other positive job attitudes (Eby et al. 2005;

Greenhaus and Powell 2006). However, most work–family

initiatives focus on the nuclear family (spouse/partner and

children). This approach is limiting as it ignores the

broader array of relationships with other family members

(e.g., parents, grandparents, adult siblings, aunts/uncles)

and friends that can anchor employees to their organiza-

tions. A nuclear family approach also takes a traditional

view that may not reflect family structures that include

single parents, childless couples, and same-sex relation-

ships (Greenhaus and Powell 2012; Özbilgin et al. 2011).

Friends may be a particularly important source of support

90 J Bus Psychol (2018) 33:89–104

123

for employees who are single or those who do not have

traditional families. Moreover, while employees may take

dependent family members with them if they move, friends

are usually left behind (Mesch and Manor 1998). Friends

could therefore play an important role in employees’

decisions to stay in their communities and workplaces.

Extending beyond the narrow confines of the nuclear

family, our study takes a more inclusive approach and

offers organizations new and important insights into the

effects of other types of family relationships, as well as

friendships, on workplace attachment. These insights can

inform managers and help organizations develop more

inclusive work–life policies and practices that retain tal-

ented workers.

Theoretical Background and Hypotheses

Here we develop the theoretical foundation for our study

by integrating the systemic model of community attach-

ment (Kasarda and Janowitz 1974) and related research

with job embeddedness theory (Mitchell et al. 2001). Our

conceptual model is displayed in Fig. 1.

The Systemic Model and the Role of Relational

Bonds in Community Attachment

The systemic model of community attachment explains the

role of relationships in community attachment. Specifi-

cally, the model holds that relationships with family and

friends root people to their communities and foster feelings

of inclusion, compatibility, belongingness, and connection

to a geographic place (Kasarda and Janowitz 1974;

Sampson 1988). Community relationships influence the

degree to which people feel ‘‘at home’’ in their community,

as well as the sorrow they would expect to experience if

they were to leave. Moreover, the model stipulates that

relationships are a greater contributor to community

attachment than the physical aspects of a geographic place

(Kasarda and Janowitz 1974; Sampson 1988). People can

find interchangeable communities with similar features, but

relationships with friends and family are not easily repli-

cated elsewhere (DaVanzo and Morrison 1981). Commu-

nity psychologists concur with the systemic model on the

importance of community relationships. These researchers

point out that people’s feelings of belongingness and

attachment to a geographic place are driven by social bonds

and their emotional connections to family and close friends

(Low and Altman 1992; Raymond et al. 2010; Scannell and

Gifford 2010).

Research on the systemic model identifies three key

aspects of relationships that connect people to a commu-

nity. First, people are attached by the proportional pres-

ence of friends and family. Proportional presence accounts

for differences in one’s family size and number of friends

(Crowe 2010; Flaherty and Brown 2010; Kasarda and

Janowitz 1974; Sampson 1988). For example, Jane and Joe

may each have three family members in their community,

but Jane may be more attached than Joe because her three

family members represent her entire family, while Joe’s

may be part of a larger family who live elsewhere. Pro-

portional presence, however, is insufficient in explaining

community attachment as it does not capture relational

quality (Beggs et al. 1996; Mesch and Manor 1998). For

example, Jane may have her entire family in her commu-

nity but may not be close to them or see them often.

Sociologists use frequency of contact as a proxy for rela-

tional quality (Belot and Ermisch 2009; Lee et al. 1990) as

Community Sacrifice

Community Fit

Workplace Turnover Intentions

Moving Intentions

Friendship Bonds

Family Bonds

Fig. 1 Community relational embeddedness model

J Bus Psychol (2018) 33:89–104 91

123

this recognizes the reciprocal influence between frequency

of interaction, affective closeness, and relationship strength

(Bengtson 2001; Homans 1951; Lawton et al. 1994). For

example, both Jane and Joe may have the same proportion

of family members in their community, but Jane sees her

family often while Joe has only infrequent interactions with

them. In addition to frequency of contact, another indicator

of relational quality used by sociologists and community

psychologists is the valence ascribed to friends and family

as a reason for staying in a community (Hidalgo and

Hernández 2001; Raymond et al. 2010; Ulrich-Schad et al.

2013). For example, Jane and Joe may both have family in

their community and see them often, but Jane may be more

likely to see family as important and the reason she stays in

her community.

Together, these three elements—proportional presence,

frequency of contact, and valence—provide a foundation

for our conceptualization of relational bonds as a key

antecedent of people’s desire to remain in their community.

Specifically, relational bonds constitute an integration of

these three elements into a single, aggregate construct

(Johnson et al. 2011). Aligned with the systemic model,

relational bonds with both family and friends would be

expected to contribute to individuals’ experience of com-

munity attachment.

Organizational and Work–Life Perspectives

on the Spillover of Community Relationships

to the Workplace

The systemic model explains how family and friendships

attach people to their communities, but does not address

the potential impact of community relationships on work-

place attachment. Work–life theory addresses the spillover

of nonwork relationships to workplace attitudes and

behavior, but fails to fully specify the mechanisms

involved. For example, Voydanoff’s (2001, 2014) model

holds that nonwork family and community relationships

create demands and resources that spill over into the work

domain. However, this model does not explain how such

life-to-work spillover occurs. Recognizing these limita-

tions, work–life scholars have called for more research and

theory that explains the mechanisms through which non-

work relationships affect work attitudes and behavior

(Greenhaus and Powell 2012; Laschober et al. 2012).

Job embeddedness theory (Lee et al. 2014; Mitchell

et al. 2001) addresses this theoretical gap by offering an

explanation of the mechanisms involved in the spillover of

nonwork experiences to workplace attachment. The theory

holds that there is a broad web of on-the-job and off-the-

job forces that determine why people stay in their organi-

zations. Although the theory offers a range of nonwork

factors that can attach employees to their organization

(e.g., home ownership, presence of a working spouse, lei-

sure activities in a community), it includes the presence of

family and friends as indicators of community links that

affect community embeddedness. More importantly, the

theory offers two constructs that can explain the mecha-

nisms underlying the relationship between community and

workplace attachment: community fit and community

sacrifice (Mitchell et al. 2001). Community fit is the degree

to which people feel compatible with their community, and

community sacrifice is people’s perception of what they

would lose if they were to leave their community. The

theory holds that people are less likely to leave their

workplaces when they have links to their community (e.g.,

presence of relationships), and when they experience

community fit and sacrifice.

Recent research has affirmed the importance of com-

munity factors in shaping workplace outcomes (Ng and

Feldman 2014). Despite this, job embeddedness research

has been limited in its approach to community relation-

ships. While researchers have examined the nuclear family,

friends are often overlooked (e.g., Lee et al. 2004; Ramesh

and Gelfand 2010). Still, two studies showed that the

friendships expatriates made abroad played a role in their

repatriation intentions (Tharenou and Caulfield 2010) and

their ability to adjust after their return (Kraimer et al.

2012). These studies, however, were specific to repatriation

and did not examine the distinct effects of friends and

family.

Related research also supports the idea that community

relationships spill over to affect organizational attachment.

Blegen, Mueller, and Price (1988), for example, found that

being responsible for family and relatives predicted

employees’ turnover intentions and organizational com-

mitment. Corporate relocation studies similarly indicate

that family can affect an employees’ willingness to relocate

for a firm (e.g., Baldridge et al. 2006; Eby and Russell

2000). However, like job embeddedness research, these

studies also have focused on the nuclear family. Important

questions thus persist concerning the role of other family

members and friends in organizational attachment.

An Integrated Model of the Spillover of Community

Relationships to Workplace Attachment

The systemic model holds that relationships with friends and

family drive community attachment because these relation-

ships determine the degree to which people feel ‘‘at home’’ in

their community, and the sorrow they would experience if

they were to leave their community (Flaherty and Brown

2010; Mesch and Manor 1998). Importantly, these constructs

are analogous to the job embeddedness constructs of

92 J Bus Psychol (2018) 33:89–104

123

community fit and sacrifice, which as defined in the pre-

ceding section, reflect the degree to which people feel

compatible with their community and people’s perception of

what they would lose if they were to leave their community

(Mitchell et al. 2001). Integrating the systemic model with

job embeddedness theory, therefore, suggests that commu-

nity fit and sacrifice are intervening mechanisms through

which relational bonds shape people’s desire to remain in a

community. As displayed in Fig. 1, relational bonds with

friends and family shape people’s feelings of community fit

and sacrifice. These perceptions in turn would be expected to

influence their desire to stay or leave their community.

Supporting this mediated relationship, scholars have found

that people express lower moving intentions (i.e., desire to

leave a community) when they feel ‘‘at home’’ in a com-

munity, and when they feel that moving would be a sacrifice

because of lost or weakened relationships (Belot and

Ermisch 2009; Guest and Lee 1983; Hidalgo and Hernández

2001; Ulrich-Schad et al. 2013). We thus hypothesize:

Hypothesis 1 The relationship between family relational

bonds and moving intentions will be mediated by

(a) community fit and (b) community sacrifice.

Hypothesis 2 The relationship between friendship rela-

tional bonds and moving intentions will be mediated by

(a) community fit and (b) community sacrifice.

We further expect that moving intentions will mediate

the influence of community fit and sacrifice on work

turnover intentions. As described above, the systemic

model supports community fit and community sacrifice as

precursors to moving intentions; however, the systemic

model does not articulate how moving intentions would

spill over to influence work-related outcomes. Job

embeddedness theory and related research suggest that

people who intend to leave their communities should also

express stronger work turnover intentions (e.g., Feldman

et al. 2012; Ragins et al. 2012). These researchers point out

that although some employees are able to retain their jobs

after leaving their communities, many would seek new

employment (Feldman et al. 2012) in order to live close to

where they work. Accordingly, this theoretical integration

suggests that community fit and sacrifice will ultimately

predict work turnover intentions through their initial

influence on moving intentions. We therefore hypothesize:

Hypothesis 3 The relationship between community fit

and work turnover intentions will be mediated by moving

intentions.

Hypothesis 4 The relationship between community sac-

rifice and work turnover intentions will be mediated by

moving intentions.

Methods

Sample and Data Collection

Our study surveyed members of a large national associa-

tion of professional accountants. A representative of the

association gave us the e-mail addresses of members who

were employed full-time and living in the USA. The use of

this national sample of accounting professionals helped us

methodologically control for geographic and regional

characteristics, which may affect moving intentions, and

for socioeconomic status, which may affect individuals’

ability to move (De Jong and Gardner 1981). In order to

increase the diversity of our sample, we oversampled for

race and ethnicity and stratified by gender. We collected

data in two rounds. We first sent an introductory e-mail to

8266 members of color who provided the association with

information about their race/ethnicity. While we are not

sure how many received this e-mail, 473 people indicated

interest and were sent an online survey link. Of these

individuals, 232 (49 %) completed the survey. We next

e-mailed 112,899 members for whom we had no infor-

mation on race/ethnicity, but selected equal proportions of

men and women in order to increase gender representation.

Of these individuals, 2859 reported interest in the research

project and were sent the survey link, with 2104 (74 %)

completing it. We sent multiple reminders and offered a

raffle of gift certificates as an incentive. In total, 70 %

(N = 2336) of those sent the link completed the survey.

However, we excluded 311 cases—243 who were not

employed in an organization, 38 who reported not having

any family or close friends, and 30 who provided largely

missing data.

Our final sample thus consisted of 2025 respondents

from the USA, with respondents coming from all 50 states.

In total, there were 1405 women and 588 men in the

sample; 1740 respondents were non-Hispanic White and

269 were people of color. More specifically, there were 107

Asian-Americans, 66 Hispanics, 50 African-Americans, 11

Native Americans, 12 multiracial individuals, and 23

reporting ‘‘other’’ (some did not disclose gender or race).

Respondents’ mean age was 42 and their mean length of

residence in their communities was 14 years. Most were

married or lived with a partner (82 %), owned their home

(89 %), and had children at home (59 %).

Measures

Unless otherwise indicated, variables were measured using

five-point Likert scales (1 = strongly disagree to

5 = strongly agree) and items were coded so that higher

J Bus Psychol (2018) 33:89–104 93

123

values represent greater amounts of the construct. All

reliability estimates exceeded .70, as displayed in Table 1.

Family Bonds and Friendship Bonds

As described earlier, relational bonds has three elements:

proportional presence, frequency of contact, and relation-

ship valence. Our measures for family bonds and friend-

ship bonds therefore comprised each of these three

elements. For family bonds, we measured proportional

presence by asking ‘‘How many of your immediate family

members live nearby?’’ while for friendship bonds we

measured proportional presence by asking ‘‘How many of

your close friends live nearby?’’ (Scale: 1 = none,

4 = most). Frequency of contact was measured with the

questions: ‘‘How often do you get together with your

immediate family?’’ and ‘‘How often do you get together

with your close friends?’’ (Scale: 1 = never, 7 = almost

every day). Finally, we measured valence for family bonds

and friendship bonds, respectively, using the statements ‘‘I

remain in this community because my family is here’’ and

‘‘I remain in this community because my close friends are

here’’ (Scale: 1 = No extent: does not affect my decision to

stay in this community, 5 = Very large extent: this is the

primary reason I stay in this community).

We used formative measurement in our family and

friendship bonds measures because the objective factors

that connect people to a community are best assessed with

cause indicators (Mitchell et al. 2001). Following current

recommendations (Diamantopoulos 2011; Johnson et al.

2011), we used structural equation modeling (SEM) and

included a reflective indicator when constructing these

formative measures. This method is superior to formative

measures that use a linear composite of equally weighted

indicators as such composite measures do not permit the

estimation of a disturbance term (Bollen and Bauldry

2011). Also, including a reflective indicator as part of the

formative measure is recommended as it helps achieve

model identification and avoid interpretational confound-

ing (Bollen 2007). As illustrated in Fig. 2, we achieved

model identification by following the ‘‘two-path emitted

rule’’ (MacKenzie et al. 2005) with a reflective path fixed

to unity (valence) and two freely estimated paths to

reflective measures (community fit and sacrifice). We

treated proportional presence and frequency of contact as

cause indicators, while we measured valence reflectively.

Community Fit and Community Sacrifice

Drawing on recent recommendations (Crossley et al. 2007;

cf., Johnson et al. 2011; Feldman et al. 2012; Ramesh and

Gelfand 2010; Zhang et al. 2012), we assessed community

fit and community sacrifice using reflective measures. In T a b le

1 D

e sc

ri p

ti v

e st

a ti

st ic

s, c o

rr e la

ti o

n s,

a n

d re

li a b

il it

ie s

V a ri

a b

le s

M e a n

S D

1 2

3 4

5 6

7 8

9 1

0 1

1 1

2 1

3 1

4

1 .

P ro

p o

rt io

n o

f fa

m il

y 2

.3 2

1 .1

5

2 .

F C

w it

h fa

m il

y 4

.7 0

1 .5

9 .6

7 *

*

3 .

F a m

il y

v a le

n c e

3 .0

1 1

.6 3

.7 3

* *

.5 9

* *

4 .

P ro

p o

rt io

n o

f fr

ie n

d s

2 .2

6 .8

0 .2

7 *

* .1

4 *

* .2

7 *

*

5 .

F C

w it

h fr

ie n

d s

4 .2

6 1

.3 3

.0 2

.0 7

* *

.0 3

.4 2

* *

6 .

F ri

e n

d sh

ip v

a le

n c e

2 .8

4 1

.3 1

.3 3

* *

.2 3

* *

.4 3

* *

.6 4

* *

.4 4

* *

7 .

L e n

g th

o f

re si

d e n

c e

1 3

.8 7

1 2

.3 6

.3 1

* *

.2 3

* *

.3 7

* *

.2 6

* *

.0 8

* *

.3 5

* *

8 .

H o

m e

o w

n e rs

h ip

.8 9

- .0

3 .0

2 .0

2 .0

1 -

.0 3

.0 2

.0 2

9 .

W o

rk in

g sp

o u

se /p

a rt

n e r

.6 6

- .0

2 .0

6 *

* .0

3 .0

3 -

.0 7

* *

- .0

3 .0

2 .1

7 *

*

1 0

. C

o m

m u

n it

y fi

t 4

.0 9

.7 2

.1 5

* *

.1 1

* *

.1 7

* *

.2 4

* *

.1 5

* *

.3 0

* *

.1 6

* *

.0 9

* *

.0 2

(. 8

6 ). 7 9

1 1

. C

o m

m u

n it

y sa

c ri

fi c e

3 .5

9 .9

6 .2

5 *

* .2

0 *

* .3

1 *

* .3

1 *

* .1

6 *

* .4

3 *

* .2

5 *

* .0

6 *

.0 3

.6 0

* *

(. 8

7 ). 8 3

1 2

. M

o v

in g

in te

n ti

o n

s 2

.1 7

.9 4

- .1

5 *

* -

.1 5

* *

- .1

9 *

* -

.1 8

* *

- .0

7 *

* -

.2 3

* *

- .1

4 *

* -

.1 0

* *

- .0

9 *

* -

.6 1

* *

- .6

0 *

* (.

8 3

). 8 0

1 3

. W

o rk

p la

c e

tu rn

o v

e r

in te

n ti

o n

s 2

.9 1

1 .7

4 -

.0 3

- .0

2 -

.0 4

- .0

6 *

* -

.0 3

- .0

8 *

* -

.1 0

* *

- .0

7 *

* -

.0 3

- .2

1 *

* -

.1 9

* *

.3 3

* *

(. 9

2 ). 9 0

1 4

. G

e n

d e r

.3 0

- .0

1 .0

1 -

.0 5

* -

.0 5

* .0

2 -

.0 3

.0 1

.0 4

* -

.1 6

* *

.0 1

.0 0

- .0

2 -

.0 3

1 5

. W

h it

e .8

7 -

.0 4

* .0

4 .0

4 .1

0 *

* .0

9 *

* .1

3 *

* .0

6 *

.0 8

* *

.0 2

.0 6

* .0

7 *

* -

.0 8

* *

- .0

5 *

- .0

4

1 6

. H

is p

a n

ic .0

3 -

.0 1

.0 2

.0 2

- .0

2 -

.0 5

* -

.0 5

* -

.0 3

- .0

2 -

.0 3

- .0

2 -

.0 1

.0 1

- .0

2 .0

4

94 J Bus Psychol (2018) 33:89–104

123

line with traditional definitions (Gusfield 1975; Puddifoot

2003), we instructed respondents to consider ‘‘community’’

as the town, city or suburb that is their primary place of

residence. We measured community fit with four items

from Mitchell et al.’s (2001) scale which have been used in

subsequent work (e.g., Lee et al. 2004): ‘‘I really love the

place where I live,’’ ‘‘I think of the community where I live

as home,’’ ‘‘This community is a good match for me,’’ and

‘‘The area where I live offers the leisure activities that I

like.’’ We measured community sacrifice with one item

from Mitchell et al.’s community sacrifice scale (‘‘Leaving

this community would be very hard’’) and two items we

developed that directly measure sacrifice: ‘‘I would give up

quite a bit by leaving this community’’ and ‘‘I would

sacrifice a lot if I left this community.’’

We explored the reliability and validity of these scales

in a separate pilot study of 96 working adults attending

graduate school at a Midwestern university. In this sample,

community fit (a = .81) and sacrifice (a = .87) offered good reliability, and preliminary convergent and discrim-

inant validity estimates were satisfactory. Community fit

and sacrifice, respectively, were positively correlated with

Buckner’s (1988) Neighborhood Cohesion scale (r = .72,

p \ .001; r = .61, p \ .001); Allen et al.’s (1991) Com- munity Satisfaction scale (r = .75, p \ .001; r = .33, p \ .01); Puddifoot’s (2003) Sense of Community Identity scale (r = .68, p \ .001; r = .54, p \ .001); and Adams’s (1992) Local Safety scale (r = .49, p \ .01; r = .27, p \ .01).

Moving Intentions and Workplace Turnover Intentions

We measured moving intentions with Ragins et al.’s

(2012) three-item Community Moving Intentions scale and

workplace turnover intentions with Cammann et al.’s

(1983) three-item Turnover Intentions scale.

Controls

In line with recommendations for the treatment of control

variables (Bernerth and Aguinis 2016; Spector and Bran-

nick 2011), we included controls that were theoretically

and empirically justified as potential confounds by job

embeddedness theory (Mitchell et al. 2001) and the sys-

temic model (Kasarda and Janowitz 1974; Sampson 1988).

Drawing on job embeddedness theory (Mitchell et al.

2001), we first controlled for two variables that affect

people’s abilities to move or change jobs: home ownership

and the presence of a working spouse or partner. In addi-

tion to being theoretically justified (Feldman et al. 2012;

Lee et al. 2014), researchers have found that owning a

home and having a working spouse link people to their

communities and affect turnover intentions and decisionsT a b le

1 c o

n ti

n u

e d

V a ri

a b

le s

M e a n

S D

1 2

3 4

5 6

7 8

9 1

0 1

1 1

2 1

3 1

4

1 7

. A

fr ic

a n

-A m

e ri

c a n

.0 3

- -

.0 5

* -

.0 3

- .0

3 -

.0 5

* -

.0 4

- .0

6 *

- .0

3 -

.0 5

* -

.0 4

- .0

4 *

- .0

5 *

.0 6

* *

.0 2

.0 3

1 8

. A

si a n

-A m

e ri

c a n

.0 5

- -

.0 4

- .0

8 *

* -

.0 6

* *

- .0

9 *

* -

.0 5

* -

.0 9

* *

- .0

4 -

.0 5

* .0

1 -

.0 3

- .0

6 *

* .0

6 *

.0 8

* *

- .0

1

N =

2 0

2 5

(N =

1 9

9 3

– 2

0 0

9 fo

r v

a ri

a b

le s

1 4

– 1

8 ). F C

fr e q

u e n

c y

o f

c o

n ta

c t.

N u

m b

e rs

o n

th e

d ia

g o

n a l

in p

a re

n th

e se

s a re

c o

e ffi

c ie

n t

a lp

h a

re li

a b

il it

ie s;

n u

m b

e rs

in it

a li

c s

a re

th e

sq u

a re

ro o

t o

f th

e

a v

e ra

g e

v a ri

a n

c e

e x

tr a c te

d .

H o

m e

o w

n e rs

h ip

is c o

d e d

1 =

o w

n s

h o

m e

a n

d 0 =

d o

e s

n o

t o

w n

h o

m e ;

w o

rk in

g sp

o u

se /p

a rt

n e r

is c o

d e d

1 =

h a s

sp o

u se

/p a rt

n e r

a n

d 0 =

d o

e s

n o

t h

a v

e

sp o

u se

/p a rt

n e r

w h

o w

o rk

s o

u ts

id e

th e

h o

m e ;

g e n

d e r

is c o

d e d

1 =

m a le

a n

d 0 =

fe m

a le

; ra

c e

a n

d e th

n ic

it y

v a ri

a b

le s

(1 5

– 1

8 )

a re

c o

d e d

, so

1 re

p re

se n

ts a n

e m

p lo

y e e

b e lo

n g

in g

to su

c h

g ro

u p

* p \

.0 5

; *

* p \

.0 1

(t w

o -t

a il

e d

te st

s)

J Bus Psychol (2018) 33:89–104 95

123

(Lee et al. 2004; Mitchell et al. 2001). We also controlled

for length of residence in the community (measured in

years) as the systemic model holds that length of residence

is an indicator of community attachment (Kasarda and

Janowitz 1974; Sampson 1988). Researchers also point out

that length of residence can confound studies of commu-

nity mobility (Flaherty and Brown 2010; Ulrich-Schad

et al. 2013). Moreover, length of residence is a common

control variable in job embeddedness research (e.g.,

Ramesh and Gelfand 2010).

Analysis Plan

Model and Hypotheses Tests

We conducted the study analyses using structural equation

modeling (SEM) in AMOS 23 with maximum likelihood

estimation. More specifically, we first conducted a confir-

matory factor analysis (CFA) of our measurement model,

which comprised all reflective study measures. We evalu-

ated the fit of the measurement model based on recognized

standards for fit statistics (Hu and Bentler 1999). We also

compared our measurement model to several nested alter-

native models using Chi-square difference tests, examined

standardized factor loadings to assess convergent validity,

and calculated the square root of the average variance

extracted for the latent variables to assess discriminant

validity (Fornell and Larcker 1981).

We next examined our hypothesized structural model.

Again, model fit was assessed based on recognized

standards for fit statistics. Given our interest in mediation,

we also compared our hypothesized structural model to five

alternative models. These tests were designed to examine

the appropriateness of the temporal order of constructs

stipulated in the hypothesized model. We used Akaike’s

Information Criterion (AIC) to conduct these tests as the

models compared were nonnested.

Finally, we used product of coefficients (Sobel’s) tests

(MacKinnon et al. 2002) to examine our hypotheses, each

of which involved mediation. To conduct these tests, we

obtained path and standard error estimates from a model

where all direct paths among latent variables were freely

estimated as this allowed for the most conservative test of

the indirect effects.

Common Method Variance (CMV)

Although we were able to obtain a large national sample of

over 2000 employees for our study, the professional asso-

ciation that gave us access to their members did not permit

us to survey their members more than once. We therefore

took proactive steps in our research design to minimize

common method bias threats. For example, as recom-

mended by Conway and Lance (2010), we took care to

demonstrate the construct validity of our measures through

multiple empirical tests, including CFAs and tests for

average variance extracted. We additionally used only

validated measures and pilot tested measures in a separate

sample if validated scales were not available. Further, we

only collected anonymous responses (Podsakoff et al.

2012). In addition to these procedural approaches, we also

Community Fit

ζ

Family Bonds

Proportion of Family Nearby

Frequency of Contact with Family

Valence

ε

Community Sacrifice

1

Fig. 2 Formative measurement of family bonds and model identification. Note a similar formative measurement was used for friendship bonds

96 J Bus Psychol (2018) 33:89–104

123

employed statistical procedures for assessing the presence

and extent of CMV in our data, which are described later.

Results

Descriptive statistics, correlations, and scale reliabilities

are reported in Table 1.

Measurement Model

The CFA of our reflective measures showed good fit

[v2(59) = 448.03, p \ .001; CFI = .98, RMSEA = .06, SRMR = .03]. This measurement model also fit better than

several alternative models, including one where the cor-

relation between community fit and sacrifice was fixed to

unity [vdiff 2

(1) = 267.75, p \ .001]. All items loaded sig- nificantly on their specified latent construct with a mean

standardized factor loading of .83, showing convergent

validity. Finally, the square root of the average variance

extracted for the latent variables, which appear on the

diagonal of the correlation matrix in Table 1, is larger than

the correlations in the same row and columns, thereby

supporting discriminant validity.

SEM Model Fit Results

Fit statistics for our hypothesized structural model also

indicated good model fit [v2(171) = 1375.73, p \ .001; CFI = .95, RMSEA = .06, SRMR = .05]. The direct

relationships are depicted in Fig. 3. Following recom-

mendations for the treatment of control variables (Spector

and Brannick 2011), we compared path estimates to a

structural model without the control variables

[v2(132) = 1075.62. CFI = .96, RMSEA = .06, SRMR = .05] and found similar results. Path coefficients

changed by no more than ±.01 and there were no changes

in statistical significance.

Structural model comparisons also indicated that our

hypothesized model had better fit to the data than several

alternative models. First, the results showed that our

hypothesized model (AIC = 1605.17) fit better than an

alternative model where workplace turnover intentions had

a direct exogenous influence on moving intentions, along

with community fit and sacrifice mediating the role of

family and friendship bonds on moving intentions

(AIC = 1692.19). Second, our hypothesized model had

better fit than a model specifying that community fit,

community sacrifice, and work turnover intentions would

mediate the role of family and friendship bonds on moving

intentions (AIC = 1684.96). These two comparisons

helped us to examine whether the data would better fit a

model where workplace turnover intentions predicted

moving intentions. Third, our model offered better fit than

a model specifying that family bonds, friendship bonds,

community fit, and community sacrifice each had a direct

exogenous influence on moving intentions, and that mov-

ing intentions mediated the effect of these four variables on

workplace turnover intentions (AIC = 1726.32). Fourth,

our model had better fit than an alternative model speci-

fying that family bonds, friendship bonds, community fit,

community sacrifice, and moving intentions all had a direct

exogenous influence on workplace turnover intentions

(AIC = 1723.30). Finally, the hypothesized model fit bet-

ter than a model in which community fit and sacrifice

mediated the effect of family and friendship bonds on

Community Sacrifice

Community Fit

Workplace Turnover Intentions

Moving Intentions

Friendship Bonds

Family Bonds

.44***

.07**

.30***

.19***

-.41***

-.49***

.43***

Fig. 3 Structural equations model with standardized path coefficient results. Note v2(171) = 1375.73, p \ .001; CFI = .95, RMSEA = .06, SRMR = .05. Standardized path coefficients

are reported. Family bonds and friendship bonds are formative

measures as depicted in Fig. 2, while the rest are reflective latent

variables. Control variables, measured items, indirect effects, and

error and disturbance terms are omitted for brevity. *p \ .05; **p \ .01; ***p \ .001

J Bus Psychol (2018) 33:89–104 97

123

moving intentions and work turnover intentions (i.e., both

moving intentions and work turnover intentions were

positioned immediately after community fit and sacrifice in

the model sequence) (AIC = 1694.93).

SEM analyses also demonstrated strong results for our

formative measures. We found significant standardized

estimates for the proportion (c = .67, p \ .001) and fre- quency of contact with family (c = .20, p \ .001) on family bonds, as well as for the proportion (c = .59, p \ .001) and frequency of contact with friends (c = .22, p \ .001) on friendship bonds.

Hypotheses Tests

Hypothesis test results revealed a significant indirect effect for

family bonds on moving intentions through both community

fit (a*b = -.01, zab = -2.62, p \ .01) and community sacrifice (a*b = -.04, zab = -6.16, p \ .001), supporting Hypotheses 1a and 1b. We also found a significant indirect

effect for friendship bonds on moving intentions through both

community fit (a*b = -.07, zab = -8.42, p \ .001) and community sacrifice (a*b = -.11, zab = -9.75, p \ .001), supporting Hypotheses 2a and 2b. Further, we found a sig-

nificant indirect effect for community fit on workplace turn-

over intentions through moving intentions (a*b = -.42, zab = -7.78, p \ .001), which supports Hypothesis 3, and a significant indirect effect for community sacrifice on work-

place turnover intentions through moving intentions

(a*b = -.41, zab = -7.88, p \ .001), which supports Hypothesis 4.

Common Method Variance

Beyond the research design strategies for mitigating the

threat of common method bias discussed above in our

analysis plan, we also assessed the presence and extent of

CMV through statistical tests. Specifically, we used the

CFA method marker technique (Williams et al. 2010) with

racial identity centrality (Sellers et al. 1998) as the

marker variable. Using this technique, we first found that

the baseline and Method-C models did not differ

[vdiff 2

(1) = 3.39, p [ .05]. However, a comparison of the Method-C and Method-U models showed the presence of a

congeneric method effect [vdiff 2

(12) = 30.11, p \ .01], and the comparison of the Method-U and Method-R models

[vdiff 2

(6) = 264.69, p \ .001] showed evidence of bias due to this congeneric method effect. We therefore assessed the

degree to which the method marker accounted for variance

in the substantive indicators, and found a median value of

only 1.8 %. In addition, reliability decomposition showed

the substantive composite reliabilities for the latent vari-

ables remained above .70, and that method reliability

ranged from .8 to 1.1 % across the constructs.

Collectively, these findings suggest that a very small, yet

statistically significant, amount of CMV may be present.

Although we found a very small effect, we retested our

hypothesized model using racial identity centrality as an

additional control variable, allowing it to have a direct

influence on all endogenous variables (Podsakoff et al.

2012). All model paths remained significant in this analy-

sis, with the standardized coefficients varying no more than

±.01. Taken together, these results suggest that while a

small degree of CMV is present, it does not appear to be a

likely explanation for our findings.

Post Hoc Analyses

We conducted two post hoc analyses. First, we examined

whether friendship bonds may have a comparatively

stronger influence in our hypothesized model than family

bonds. Second, we examined if our model held across

different racial, ethnic, and gender groups.

Comparative Tests for Friendship and Family Bonds

An interesting pattern of findings in our SEM analyses and

mediation tests was that effects for friendship bonds on

community fit, sacrifice, and moving intentions were

notably stronger than effects for family bonds. Given this,

we compared our proposed model to models where the

effects of family and friendship bonds were fixed to be

equal. Results showed that the effects of friendship bonds

on community fit [vdiff 2

(1) = 29.90, p \ .001] and on community sacrifice [vdiff

2 (1) = 34.36, p \ .001] were

significantly stronger than the effect of family bonds.

Taken together, these results suggest that friendships could

ultimately play a more integral role than family in shaping

individuals’ community and workplace attachment.

The Role of Diversity in Community and Workplace

Attachment

Diversity scholars point out that because race, gender, and

ethnicity create different life experiences, researchers

should not assume diversity does not matter, but instead

examine if effects hold across diverse groups (Roberson

et al., in press). Following this recommendation, we con-

ducted multigroup SEM analyses comparing women and

men, as well as comparing Whites, Hispanics, African-

Americans and Asian-Americans. Table 2 summarizes

these results, which offered several interesting findings. For

example, although it is traditionally assumed that rela-

tionships are more important for women (Cross and Mad-

son 1997; Swartz 2009), our results showed that men and

women did not differ in the impact of family bonds on

community fit or sacrifice. In addition, friendship bonds,

98 J Bus Psychol (2018) 33:89–104

123

while important for both men and women, were in fact a

stronger predictor of community fit for men than for

women.

As displayed in Table 2, several interesting effects

emerged that illustrated racial and ethnic differences in the

effects of friends and family on community embeddedness.

For example, family played an important role in Hispanics’

experiences of community sacrifice (i.e., what they would

lose if they left their community) that was significantly

stronger than all other racial groups (i.e., Whites, African-

Americans and Asian-Americans). However, Hispanics’

experiences of community fit, which reflects their feelings

of ‘‘fitting in’’ with their community, was driven more by

their friends than by their family. Although it is frequently

assumed that family is a binding force for Hispanics, this

analysis, while exploratory, suggests that although family

may be important for Hispanics (Freeberg and Stein 1996;

Mindel 1980), the effects of family on their experiences of

community embeddedness may be quite nuanced and vary

by experiences of fit and sacrifice.

Discussion

This study examined the spillover of community relation-

ships to the workplace. Broadening the research lens, we

went beyond the nuclear family to examine how employ-

ees’ relationships with other family members and friends

spill over to affect organizational attachment. By inte-

grating the systemic model of community attachment

(Kasarda and Janowitz 1974; Sampson 1988) and job

embeddedness theory (Mitchell et al. 2001), we unpacked

and studied the mechanisms underlying these spillover

effects. We also applied theory and research from the

sociological and community psychology literatures (e.g.,

Belot and Ermisch 2009; Raymond et al. 2010) to capture

the proportional presence and quality of employees’ com-

munity relationships.

Using survey data from a national sample of 2025 pro-

fessionals, we found support for our conceptual model.

Employees’ bonds with friends and family predicted their

attachment to their communities, which in turn predicted

their workplace turnover intentions. We also found that

community fit and sacrifice were important mediating

mechanisms that explained these relationships. Specifi-

cally, bonds with friends and family predicted employees’

intentions to leave their communities, and this association

was mediated by employees’ experiences of fit with their

community, and by their perceptions of the sacrifice they

would experience if they were to leave their community. In

line with past job embeddedness research (Jiang et al.

2012; Mitchell et al. 2001; Ragins et al. 2012), those

experiencing stronger community fit and sacrifice also had

lower workplace turnover intentions, with moving inten-

tions mediating these relationships. Taken together, these

findings support the idea that nonwork relationships play a

key role in connecting employees to their communities and

workplaces.

A key finding of our study was that friendships played a

central role in anchoring employees to their communities

and ultimately their workplaces. In fact, friends had a

significantly stronger impact on community attachment

than family. These findings are in line with psychological

and sociological research showing that friends offer unique

Table 2 Multigroup structural equations model analyses results

Structural path Gender Race/ethnic group

Males Females Whites Hispanics African-Americans Asian-Americans

Family bonds ? community fit .06 .06* .08** -.24*

-2.45*

.05

2.03*

-.04

Family bonds ? community sacrifice .22*** .16*** .18*** .44**

2.36*

.19

22.27*

.05

22.36*

Friendship bonds ? community fit .39*** .28***

-1.96*

.29*** .56***

2.67**

.26

2.2.91**

.36**

Friendship bonds ? community sacrifice .43*** .42*** .41*** .44*** .39* .40**

Community fit ? moving intentions -.30*** -.44***

-2.19*

-.38*** -.68***

-2.41*

-.55** -.65***

-2.23*

Community sacrifice ? moving intentions -.61*** -.44*** -.50*** -.34 -.20 -.30**

Moving intentions ? workplace turnover intentions .45*** .41*** .42*** .62*** .28 .56***

Standardized coefficients are reported. Z scores for significant group differences are reported beneath the coefficient of the second group. Z scores

in italics involve Whites as the referent group, and Z scores in bold involve Hispanics as the referent group

* p \ .05; ** p \ .01; *** p \ .001

J Bus Psychol (2018) 33:89–104 99

123

forms of social support that differ from the ones provided

by family (Baumeister and Leary 1995; House et al. 1988).

Family relationships are rooted in responsibility and obli-

gation (Blegen et al. 1988; Lee et al. 1990), while friend-

ships are voluntary and chosen to fill affiliation, social

connection, and belongingness needs (Wellman and

Wortley 1990; Wiseman 1986). Friends are also more

likely than family to be lost after leaving a community

(Belot and Ermisch 2009; Gillespie et al. 2015; Mesch and

Manor 1998). This research may explain the strong rela-

tionships between friendships and community fit and sac-

rifice we found in the present study. Friends provide a

powerful source of fit with a community, and since friends

usually do not relocate with employees, they also reflect a

strong source of sacrifice. Although friendships are typi-

cally overlooked in work–life research (cf., Eby et al.

2005), our findings suggest that scholars should incorporate

friendships into their research, and that organizations

should recognize the role of friends in their policies and

programs.

As described later, these findings have significant

implications for management practitioners and organiza-

tions. For example, organizations that require relocation

and offer employees the ‘‘benefit’’ of assisting with relo-

cating nuclear families may miss ‘‘the big picture’’ by

ignoring the role of other family members and friends in

employees’ attachment to their communities and

workplaces.

Implications for Theory and Future Research

Our results highlight the importance of community rela-

tionships in organizational life. Sociologists have long

observed that people are connected to their communities

through a constellation of relationships (Kasarda and

Janowitz 1974; Sampson 1988), but both work–life (cf.,

Eby et al. 2005) and job embeddedness (e.g., Lee et al.

2004) research has focused primarily on the nuclear family.

We extended the scope of study to incorporate family

relationships beyond the nuclear family, as well as

friendships. Our findings indicate that friends may embed

employees to their communities and workplaces in ways

that are distinct and often stronger than some family

relationships.

In fact, we found that while family is important, rela-

tionships with friends had a stronger impact on community

and workplace attachment. Although this counters some

forms of proverbial wisdom (e.g., ‘‘blood is thicker than

water’’ and ‘‘an ounce of blood is worth more than a pound

of friendship’’), it is aligned with others (e.g., ‘‘Friends are

God’s apology for relations’’—Hugh Kingsmill) and the

fact that we can choose our friends, but not our family.

These findings indicate that scholars should include non-

work friendships and nonnuclear family relationships in

their assessment of organizational attachment and the

work–life interface.

Our study also offers scholars a more complete approach

to the assessment of community relationships. Drawing on

the sociological and community psychology literatures, we

developed the construct of relational bonds, which incor-

porates not only the proportional presence of relationships

in a community, but also relational quality. Scholars may

use this construct and similar measurement to more fully

capture the spillover of family and friendship relationships

to workplace attitudes and behavior, and to incorporate the

important role of relationship quality in employee’s expe-

riences. We operationalized relational quality as frequency

of contact and relationship valence. Future research may

incorporate other aspects or measures of relationship

quality, such as affective connection or caring, and provide

a more systemic comparison of the impact of different

types of nonnuclear family relationships on workplace

experiences. Scholars could also examine the spillover

effects of community relationships stemming from places

of worship, volunteer and neighborhood activities on

workplace experiences and attachment.

We focused on workplace turnover intentions in this

study, but future research could also extend our analysis to

examine the impact of community relationships on other

work experiences and outcomes. Scholars could use the life

spillover model to examine the spillover of positive com-

munity relationships to work attitudes and behaviors (Ra-

gins et al. 2014). Positive life spillovers are nonwork

experiences that build employees’ psychological resources

and capacity. Applying this model, researchers could

replicate and extend our findings by examining whether

positive community relationships spill over to affect

employees’ well-being at work, as well as other work

attitudes and behaviors such as job satisfaction and orga-

nizational citizenship behaviors. Positive community rela-

tionships could also provide employees with psychological

resources that build their resilience for work stress.

Lastly, although limited by small samples, our post hoc

analyses involving diversity showed intriguing differences

that can be a springboard for future research, particularly

on the role of friendships in community and workplace

experiences. For example, in contrast to traditional

expectations, we found that friendships played a stronger

role in experiences of community fit for men than for

women. We also found that friendships played a particu-

larly strong role in the experiences of community fit for

Hispanics. Future research could examine gender, racial,

and ethnic differences in the types of behaviors, roles, and

functions in friendships that create these psychological

experiences.

100 J Bus Psychol (2018) 33:89–104

123

Implications for Practice

Our findings have a number of important practical impli-

cations for managers, practitioners, human resource pro-

fessionals, and organizations. To start, organizations

should consider taking a broader approach in defining

nonwork relationships by including friends, and by

extending the types of family relationships included in

work–life programs, policies and practices. For example,

firms typically give employees time off to care for a sick

child or spouse, but few give comparable accommodations

for friends or relatives. Taking a flexible approach to

‘‘family days’’ (e.g., ‘‘personal days’’) may be one way to

address this limitation. As another example, most organi-

zations support the nuclear family with on-site childcare or

childcare copayments, but many fail to include other

family members in their dependent care policies.

Overall, our findings suggest that organizations need to

recognize the valuable role nonwork friendships play in

workers’ lives. For example, friends may be particularly

important for those who, for a variety of reasons, cannot

rely on a spouse or partner for support. Friends often fall

outside the realm of most work–life policies and programs,

but consistent with an enrichment approach to work–life

(Greenhaus and Powell 2006; Voydanoff 2014), may be an

important resource for employees. In fact, in many ways,

friendships may be less demanding than family. While

family is often a source of normative obligations, time

constraints, energy demands, and work–family conflict

(Feldman et al. 2012; Greenhaus and Powell 2012),

friendships are often more flexible and can be a potent

source of companionship, advice and support for employ-

ees (Fehr 1996; Wiseman 1986). It is also important to

keep in mind that friendships do need time to be developed,

nurtured, and sustained, and employees need to be able to

fulfill their roles and responsibilities for their friends as

well as their families. Organizations can allow their

employees opportunities to foster strong friendship bonds

and community attachment by encouraging them to form,

maintain, and strengthen their community relationships.

For example, organizations can promote and reward par-

ticipation in volunteer and community activities that can

help their employees develop close personal relationships

outside the workplace.

Our findings also suggest that organizations may need to

reconsider relocation as a career strategy for their

employees. While some employees may agree to relocate

to get a promotion or a raise, the subsequent loss of close

friends and the strain on family relationships could ulti-

mately undermine the employee’s organizational attach-

ment and commitment. In contrast, the presence of strong

friends and family bonds in a community can anchor

employees and help retain them in the organization.

In short, organizations should recognize that their

employees receive emotional and instrumental support and

resources from relationships that extend beyond the nuclear

family. A broad and inclusive conception of relationships

may help firms reap the rewards of workplace attachment,

which in turn may lead to other desired outcomes such as

lower absenteeism and counterproductive behaviors, as

well as greater citizenship behaviors and role performance

(Kiazad et al. 2015; Lee et al. 2004; Mai et al. 2016).

Limitations

Our study has limitations that need to be acknowledged.

First, our cross-sectional research design did not allow us

to make causal inferences about the directions of the

relationships studied. We were able, however, to take

several steps to mitigate the threat of method bias (Conway

and Lance 2010), including implementing a research

design that incorporated objective measures for propor-

tional presence of family and friends in the community and

frequency of contact with them. Our statistical CMV tests

also revealed that although there was congeneric CMV

present, the extent of this CMV was small and not a likely

explanation for our findings. Importantly, CMV also cannot

account for group differences (Siemsen et al. 2010) or our

finding that friends had a stronger impact on community

and organizational attachment than family.

Another potential limitation involves the directions of

the relationships studied. We examined the spillover of

community relationships to the work domain (i.e., life-to-

work spillover), but workplace experiences can also spill

over to the life domain (i.e., work-to-life spillover). As

described in our results section, we compared our model

with five alternative structural models, including two

models with a reversed causal path in which workplace

turnover intentions predict community moving intentions.

Our model had better fit than all of the alternatives.

However, future research could use longitudinal designs to

examine how the relationship between turnover and mov-

ing intentions may unfold over time.

Our study had a few other potential limitations. First, we

examined moving and turnover intentions rather than

behaviors. However, moving intentions are predictors of

behaviors and migration (De Jong and Gardner 1981; Lee

et al. 1994; Mitchell et al. 2001) and are not subject to

financial constraints that prevent people from leaving a job

or community (Lu 2011; Ulrich-Schad et al. 2013). Second,

in order to increase gender diversity, we sampled equal

numbers of male and female members of the professional

association of accountants. However, women were more

likely to respond to the survey than men, perhaps reflecting

a response bias. Although we were unable to compare

J Bus Psychol (2018) 33:89–104 101

123

respondents with those who chose not to participate in the

study, our sample closely mirrored the industry and spe-

cializations of the members of the professional association

we surveyed. Still, our sample may differ from the larger

population of the professional association and the profes-

sion, which may limit the generalizability of the results.

Lastly, while our national sample of professionals in the

USA allowed us to control for socioeconomic status and

regional and community characteristics that may influence

moving intentions, the results may not generalize to blue-

collar occupations or employees outside the USA. We hope

that future research extends our study by examining com-

munity relational embeddedness in other occupations and

countries. We further encourage future research to replicate

and extend our findings using more work and organiza-

tional variables and outcomes.

Conclusion

The findings of our study reinforce emerging theory on the

role of relationships in employees’ lives (Eby and Allen

2012; Dutton and Ragins 2007; Lieberman 2013) and

support the tenet that relationships are the means through

which employees become connected to their communities

and workplaces. Friends and family are not only funda-

mental relationships driven by the basic human need to

form and maintain interpersonal bonds (Baumeister and

Leary 1995), they also are powerful sources of attachment

that anchor employees to their communities and work-

places. Our examination offers new and important insights

into the spillover of friendships and nonnuclear family to

the workplace, and suggests that scholars and managers

consider a wide range of employee relationships in the

work–life interface. In so doing, they should not overlook

the value of close friends and a broad array of family

relationships.

References

Adams, R. (1992). Is happiness a home in the suburbs? The influence

of urban versus suburban neighborhoods on psychological

health. Journal of Community Psychology, 20, 353–372.

Allen, L., Long, P., & Perdue, R. (1991). Relational patterns between

community dimensions and global measures of community

satisfaction. Journal of Rural Studies, 7, 331–338.

Baldridge, D. C., Eddleston, K. A., & Veiga, J. F. (2006). Saying no

to being uprooted: The impact of family and gender on

willingness to relocate. Journal of Occupational and Organiza-

tional Psychology, 79, 131–149.

Baumeister, R. F., & Leary, M. R. (1995). The need to belong: Desire

for interpersonal attachments as a fundamental human motiva-

tion. Psychological Bulletin, 117, 497–529.

Beggs, J. J., Hurlbert, J. S., & Haines, V. A. (1996). Community

attachment in a rural setting: A refinement and empirical test of

the systemic model. Rural Sociology, 61, 407–426.

Belot, M., & Ermisch, J. (2009). Friendship ties and geographical

mobility: Evidence from Great Britain. Journal of the Royal

Statistical Society: Series A (Statistics in Society), 172, 427–442.

Bengtson, V. L. (2001). Beyond the nuclear family: The increasing

importance of multigenerational bonds. Journal of Marriage and

Family, 63, 1–16.

Bernerth, J. B., & Aguinis, H. (2016). A critical review and best-

practice recommendations for control variable usage. Personnel

Psychology, 69, 229–283.

Blegen, M. A., Mueller, C. W., & Price, J. L. (1988). Measurement of

kinship responsibility for organizational research. Journal of

Applied Psychology, 73, 402–409.

Bollen, K. A. (2007). Interpretational confounding is due to

misspecification, not to type of indicator: Comment on Howell,

Breivik, and Wilcox (2007). Psychological Methods, 12,

219–228.

Bollen, K. A., & Bauldry, S. (2011). Three Cs in measurement

models: Causal indicators, composite indicators, and covariates.

Psychological Methods, 16, 265–284.

Buckner, J. C. (1988). The development of an instrument to measure

neighborhood cohesion. American Journal of Community Psy-

chology, 16, 771–791.

Cammann, C., Fichman, M., Jenkins, D., & Klesh, J. (1983).

Assessing the attitudes and perceptions of organizational mem-

bers. In S. Seashore, E. Lawler, P. Mirvis, & C. Cammann

(Eds.), Assessing organizational change: A guide to methods,

measures and practices (pp. 71–138). New York: Wiley.

Carmeli, A., & Gittell, J. H. (2009). High-quality relationships,

psychological safety, and learning from failures in work

organizations. Journal of Organizational Behavior, 30, 709–729.

Casper, W. J., Eby, L. T., Bordeaux, C., Lockwood, A., & Lambert,

D. (2007). A review of research methods in IO/OB work–family

research. Journal of Applied Psychology, 92, 28–43.

Conway, J. M., & Lance, C. E. (2010). What reviewers should expect

from authors regarding common method bias in organizational

research. Journal of Business and Psychology, 25, 325–334.

Cross, S. E., & Madson, L. (1997). Models of the self: Self-construals

and gender. Psychological Bulletin, 122, 5–37.

Crossley, C. D., Bennett, R. J., Jex, S. M., & Burnfield, J. L. (2007).

Development of a global measure of job embeddedness and

integration into a traditional model of voluntary turnover.

Journal of Applied Psychology, 92, 1031–1042.

Crowe, J. (2010). Community attachment and satisfaction: The role of

a community’s social network structure. Journal of Community

Psychology, 38, 622–644.

DaVanzo, J. S., & Morrison, P. A. (1981). Return and other sequences

of migration in the United States. Demography, 18, 85–101.

De Jong, G. F., & Gardner, R. W. (Eds.). (1981). Migration decision

making: Multidisciplinary approaches to microlevel studies in

developed and developing countries. New York: Pergamon.

Diamantopoulos, A. (2011). Incorporating formative measures into

covariance-based structural equation models. MIS Quarterly, 35,

335–358.

Dutton, J. E., & Ragins, B. R. (Eds.). (2007). Exploring positive

relationships at work: Building a theoretical and research

foundation. Mahwah, NJ: Lawrence Erlbaum.

Eby, L. T., & Allen, T. D. (Eds.). (2012). Personal relationships: The

effect on employee attitudes, behavior, and well-being. New

York: Routledge.

Eby, L. T., Casper, W. J., Lockwood, A., Bordeaux, C., & Brinley, A.

(2005). Work and family research in IO/OB: Content analysis

and review of the literature (1980–2002). Journal of Vocational

Behavior, 66, 124–197.

Eby, L. T., & Russell, J. E. (2000). Predictors of employee

willingness to relocate for the firm. Journal of Vocational

Behavior, 57, 42–61.

102 J Bus Psychol (2018) 33:89–104

123

Fehr, B. (1996). Friendship processes. Thousand Oaks, CA: Sage.

Feldman, D. C., Ng, T. W., & Vogel, R. M. (2012). Off-the-job

embeddedness: A reconceptualization and agenda for future

research. In J. J. Martocchio, A. Joshi, & H. Liao (Eds.),

Research in personnel and human resources management (Vol.

31, pp. 209–251). Bingley: Emerald.

Flaherty, J., & Brown, R. B. (2010). A multilevel systemic model of

community attachment: Assessing the relative importance of the

community and individual levels. American Journal of Sociol-

ogy, 116, 503–542.

Fornell, C., & Larcker, D. F. (1981). Structural equation models with

unobservable variables and measurement error: Algebra and

statistics. Journal of Marketing Research, 18, 382–388.

Freeberg, A. L., & Stein, C. H. (1996). Felt obligation towards parents

in Mexican-American and Anglo-American young adults. Jour-

nal of Social and Personal Relationships, 13, 457–471.

Gillespie, B. J., Lever, J., Frederick, D., & Royce, T. (2015). Close

adult friendships, gender, and the life cycle. Journal of Social

and Personal Relationships, 32, 709–736.

Greenhaus, J. H., & Powell, G. N. (2006). When work and family are

allies: A theory of work–family enrichment. Academy of

Management Review, 31, 72–92.

Greenhaus, J. H., & Powell, G. N. (2012). The family-relatedness of

work decisions: A framework and agenda for theory and

research. Journal of Vocational Behavior, 80, 246–255.

Guest, A. M., & Lee, B. A. (1983). Sentiment and evaluation as

ecological variables. Sociological Perspectives, 26, 159–184.

Gusfield, J. R. (1975). Community: A critical response. New York:

Harper & Row.

Hidalgo, M. C., & Hernández, B. (2001). Place attachment: Concep-

tual and empirical questions. Journal of Environmental Psy-

chology, 21, 273–281.

Homans, G. C. (1951). The human group. New York: Routledge.

House, J. S., Umberson, D., & Landis, K. R. (1988). Structures and

processes of social support. Annual Review of Sociology, 14,

293–318.

Hu, L., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in

covariance structure analysis: Conventional criteria versus new

alternatives. Structural Equation Modeling: A Multidisciplinary

Journal, 6, 1–55.

Jiang, K., Liu, D., McKay, P. F., Lee, T. W., & Mitchell, T. R. (2012).

When and how is job embeddedness predictive of turnover? A

meta-analytic investigation. Journal of Applied Psychology, 97,

1077–1096.

Johnson, R. E., Rosen, C. C., & Chang, C. H. (2011). To aggregate or

not to aggregate: Steps for developing and validating higher-

order multidimensional constructs. Journal of Business and

Psychology, 26, 241–248.

Kasarda, J. D., & Janowitz, M. (1974). Community attachment in

mass society. American Sociological Review, 39, 328–339.

Kiazad, K., Holtom, B. C., Hom, P. W., & Newman, A. (2015). Job

embeddedness: A multifoci theoretical extension. Journal of

Applied Psychology, 3, 641–659.

Kraimer, M. L., Shaffer, M. A., Harrison, D. A., & Ren, H. (2012).

No place like home? An identity strain perspective on repatriate

turnover. Academy of Management Journal, 55, 399–420.

Laschober, T. C., Allen, T. D., & Eby, L. T. (2012). Negative

nonwork relational exchanges and links to employees’ work

attitudes, work behaviors, and well-being. In L. T. Eby & T.

D. Allen (Eds.), Personal relationships: The effect on employee

attitudes, behavior, and well-being (pp. 325–348). New York:

Routledge.

Lawton, L., Silverstein, M., & Bengtson, V. (1994). Affection, social

contact, and geographic distance between adult children and their parents. Journal of Marriage and the Family, 56, 57–68.

Lee, T. W., Burch, T. C., & Mitchell, T. R. (2014). The story of why

we stay: A review of job embeddedness. Annual Review of

Organizational Psychology and Organizational Behavior, 1,

1–18.

Lee, T. R., Mancini, J. A., & Maxwell, J. W. (1990). Sibling

relationships in adulthood: Contact patterns and motivations.

Journal of Marriage and Family, 52, 431–440.

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

B. C. (2004). The effects of job embeddedness on organizational

citizenship, job performance, volitional absences, and voluntary

turnover. Academy of Management Journal, 47, 711–722.

Lee, B. A., Oropesa, R. S., & Kanan, J. W. (1994). Neighborhood

context and residential mobility. Demography, 31, 249–270.

Lieberman, M. D. (2013). Social: Why our brains are wired to

connect. New York: Crown.

Low, S. M., & Altman, I. (1992). Place attachment: A conceptual

inquiry. In I. Altman & S. M. Low (Eds.), Place attachment (pp.

1–12). New York: Plenum Press.

Lu, M. (2011). Analyzing migration decision making: Relationships

between residential satisfaction, mobility intentions, and moving

behavior. Environment and Planning A, 30, 1473–1495.

MacKenzie, S. B., Podsakoff, P. M., & Jarvis, C. B. (2005). The

problem of measurement model misspecification in behavioral

and organizational research and some recommended solutions.

Journal of Applied Psychology, 90, 710–730.

MacKinnon, D. P., Lockwood, C. M., Hoffman, J. M., West, S. G., &

Sheets, V. (2002). A comparison of methods to test mediation

and other intervening variable effects. Psychological Methods, 7,

83–104.

Mai, K. M., Ellis, A. P., Christian, J. S., & Porter, C. O. (2016).

Examining the effects of turnover intentions on organizational

citizenship behaviors and deviance behaviors: A psychological

contract approach. Journal of Applied Psychology,. doi:10.1037/

apl0000115.

Mesch, G. S., & Manor, O. (1998). Social ties, environmental

perception, and local attachment. Environment and Behavior, 30,

504–519.

Mindel, C. H. (1980). Extended familism among urban Mexican

Americans, Anglos, and Blacks. Hispanic Journal of Behavioral

Sciences, 2, 21–34.

Mitchell, T. R., Holtom, B. C., Lee, T. W., Sablynski, C. J., & Erez,

M. (2001). Why people stay: Using job embeddedness to predict

voluntary turnover. Academy of Management Journal, 44,

1102–1121.

Ng, T. W., & Feldman, D. C. (2014). Community embeddedness and

work outcomes: The mediating role of organizational embed-

dedness. Human Relations, 67, 71–103.

Özbilgin, M. F., Beauregard, T. A., Tatli, A., & Bell, M. P. (2011).

Work–life, diversity and intersectionality: A critical review and

research agenda. International Journal of Management Reviews,

13, 177–198.

Podsakoff, P. M., MacKenzie, S. B., & Podsakoff, N. P. (2012).

Sources of method bias in social science research and recom-

mendations on how to control it. Annual Review of Psychology,

65, 539–569.

Procidano, M. E., & Heller, K. (1983). Measures of perceived social

support from friends and from family: Three validation studies.

American Journal of Community Psychology, 11, 1–24.

Puddifoot, J. E. (2003). Exploring ‘‘personal’’ and ‘‘shared’’ sense of

community identity in Durham city, England. Journal of

Community Psychology, 31, 87–106.

Ragins, B. R., Gonzalez, J. A., Ehrhardt, K., & Singh, R. (2012).

Crossing the threshold: The spillover of community racial

diversity and diversity climate to the workplace. Personnel Psychology, 65, 755–787.

J Bus Psychol (2018) 33:89–104 103

123

Ragins, B. R., Lyness, K., Williams, L., & Winkel, D. (2014). Life

spillovers: The spillover of fear of home foreclosure to the

workplace. Personnel Psychology, 67, 763–800.

Ramesh, A., & Gelfand, M. J. (2010). Will they stay or will they go?

The role of job embeddedness in predicting turnover in

individualistic and collectivistic cultures. Journal of Applied

Psychology, 95, 807–823.

Raymond, C. M., Brown, G., & Weber, D. (2010). The measurement

of place attachment: Personal, community, and environmental

connections. Journal of Environmental Psychology, 30,

422–434.

Roberson, Q., Ryan, A. M. & Ragins, B. R. (in press). The evolution

and future of diversity at work. Journal of Applied Psychology.

Sampson, R. J. (1988). Local friendship ties and community

attachment in mass society: A multilevel systemic model.

American Sociological Review, 53, 766–779.

Scannell, L., & Gifford, R. (2010). Defining place attachment: A

tripartite organizing framework. Journal of Environmental

Psychology, 30, 1–10.

Sellers, R. M., Smith, M. A., Nicole Shelton, J., Rowley, S. A., &

Chavous, T. M. (1998). Multidimensional model of racial

identity: A reconceptualization of African American racial

identity. Personality and Social Psychology Review, 2, 18–39.

Siemsen, E., Roth, A., & Oliveira, P. (2010). Common method bias in

regression models with linear, quadratic, and interaction effects.

Organizational Research Methods, 13, 456–476.

Spector, P. E., & Brannick, M. T. (2011). Methodological urban

legends: The misuse of statistical control variables. Organiza-

tional Research Methods, 14, 287–305.

Stephens, J. P., Heaphy, E. D., Carmeli, A., Spreitzer, G. M., &

Dutton, J. E. (2013). Relationship quality and virtuousness:

Emotional carrying capacity as a source of individual and team

resilience. The Journal of Applied Behavioral Science, 9, 13–41.

Swartz, T. T. (2009). Intergenerational family relations in adulthood:

Patterns, variations, and implications in the contemporary United

States. Annual Review of Sociology, 35, 191–212.

Tharenou, P., & Caulfield, N. (2010). Will I stay or will I go?

Explaining repatriation by self-initiated expatriates. Academy of

Management Journal, 53, 1009–1028.

Ulrich-Schad, J. D., Henly, M., & Safford, T. G. (2013). The role of

community assessments, place, and the great recession in the

migration intentions of rural Americans. Rural Sociology, 78,

371–398.

Voydanoff, P. (2001). Incorporating community into work and family

research: A review of basic relationships. Human Relations, 54,

1609–1637.

Voydanoff, P. (2014). Work, family, and community: Exploring

interconnections (2nd ed.). New York: Psychology Press.

Wellman, B., & Wortley, S. (1990). Different strokes from different

folks: Community ties and social support. American Journal of

Sociology, 96, 558–588.

Williams, L. J., Hartman, N., & Cavazotte, F. (2010). Method

variance and marker variables: A review and comprehensive

CFA marker technique. Organizational Research Methods, 13,

477–514.

Wiseman, J. P. (1986). Friendship: Bonds and binds in a voluntary

relationship. Journal of Social and Personal Relationships, 3,

191–211.

Zhang, M., Fried, D. D., & Griffeth, R. W. (2012). A review of job

embeddedness: Conceptual, measurement issues, and directions

for future research. Human Resource Management Review, 22,

220–231.

104 J Bus Psychol (2018) 33:89–104

123